Study on the Factors Influencing the Adoption of Intelligent Agricultural Machinery by Farmers in Changsha County, Hunan Province, Based on the Ordered Logit Model

Abstract

In order to better promote the use of intelligent agricultural machinery, enhance the efficiency of grain production, optimize resource utilization, and effectively address the practical problem of the reduction in the rural labor force, while theoretically clarifying the mechanism that affects the adoption of intelligent agricultural machinery by farmers in Changsha County. Based on a questionnaire survey of farmers in Changsha County, Hunan Province, the ordered logit model was used to identify the significant factors influencing farmers’ adoption of intelligent agricultural machinery. The empirical results show that male farmers, farmers with a non-agricultural occupation, and farmers with a lower education level (below high school) have a lower willingness to adopt intelligent agricultural machinery. As the risk of purchasing intelligent agricultural machinery decreases, market demand increases, and the number of agricultural services provided by the government increases, the likelihood of farmers adopting intelligent agricultural machinery also increases. Based on these findings, this paper proposes targeted suggestions aimed at increasing the adoption of intelligent agricultural machinery by farmers in Changsha County, Hunan Province.

1. Introduction

Food is the foundation of life, and food safety is a top priority. Food is not only the material basis for society’s survival and development, but also a vital strategic resource for national security [1]. Ensuring food security for 1.4 billion people has always been the primary task of national governance. We must fully recognize the importance of food security from the perspective of national security and the overall interests of the people, and remain highly vigilant at all times. On 30 May 2020, General Secretary Xi Jinping proposed enhancing grain production capacity through scientific and technological innovation, the promotion of modern agricultural technologies, and the development of intelligent equipment. In the “Guiding Opinions on Vigorously Developing Intelligent Agriculture” issued in 2024, the Ministry of Agriculture and Rural Affairs explicitly emphasized that efforts should be based on the actual conditions of China’s agricultural and rural development to promote the optimization and upgrading of agricultural machinery [2].

Furthermore, China has a large population; although urbanization is accelerating, the rural population still accounts for a significant proportion. In addition, most farming households still operate on a family basis, with fragmented and small-scale landholdings; they generally lack advanced agricultural technology, resulting in relatively low production efficiency. Consequently, smallholder farming remains the predominant agricultural model in China, reflecting the fundamental reality of China as a “country of smallholders.” Although large-scale agricultural operations have made some progress in recent years, individual farming households remain the mainstay of agricultural production. According to statistics, more than 60% of arable land is still cultivated by individual farming households, who produce grains, vegetables, and fruits on their own plots to meet the demands of urban and rural markets. Consequently, farming households play a vital role in agricultural production and exert an irreplaceable influence on the preservation of agricultural traditions, the stability of agricultural output, farmers’ employment and income growth, and the harmony of rural society [3]. Therefore, identifying the factors influencing farmers in Changsha County’s adoption of intelligent agricultural machinery will help stimulate their enthusiasm for purchasing such equipment, promote its wider application, and thereby significantly improve grain production efficiency and resource utilization, effectively addressing the practical challenges posed by the decline in the rural labor force [4].

Intelligent agricultural machinery refers to modern equipment that integrates advanced technologies such as the Internet of Things (IoT), artificial intelligence (AI), Global Positioning System (GPS), automated control, and big data analytics, and is capable of environmental sensing, autonomous decision-making, precision operations, and remote monitoring and control [5]. Intelligent agricultural machinery not only performs the mechanical functions of traditional machinery—replacing manual labor and improving operational efficiency—but also emphasizes the automation and precision of agricultural production processes through data-driven and intelligent decision-making [6].

Scholars both in China and abroad have conducted detailed studies on the factors influencing farmers’ adoption of intelligent agricultural machinery. DJIBO found that farmers’ adoption of different technologies is primarily influenced by household economic conditions, resource endowments, and regional environments. Farmers with better economic conditions are more willing to invest in new equipment and technologies, and those with ample resources are more likely to adopt high-cost equipment. Meanwhile, regional environments such as soil and climate, as well as government policy support, directly affect the penetration rate of agricultural machinery [7]; Yichun Li found that rice farmers receiving more subsidies are more willing to adopt intelligent agricultural machinery [8]. Ni Yang conducted a survey in Yiyang, Hunan, and found that among the 12 influencing factors he listed, the farmers interviewed placed the highest importance on training for the use of intelligent agricultural machinery. Many farmers stated that intelligent agricultural machinery is practical, but they are concerned about not knowing how to operate it, the difficulty of training, and the high maintenance costs that follow [9]. Nie Zhengyan’s research indicates that older farmers view intelligent agricultural machinery as high-end technology and find it difficult to trust its reliability; compared to younger farmers, this group has a relatively lower willingness to adopt such technology [10]. Meanwhile, Liu Jie’s research also indicates that, compared to older farmers, younger farmers show greater interest in intelligent agricultural machinery and are more likely to adopt new intelligent devices [11]. Wang Yanjian’s related research shows that the higher the educational level, the more exposure to emerging technologies, and the higher the acceptance of new intelligent agricultural machinery [12]. Gai Hao’s research found that households with better economic conditions and lower housing-related financial burdens face less financial pressure when purchasing new agricultural machinery and thus exhibit a higher willingness to buy new agricultural machinery and adopt emerging technologies. In contrast, small-scale farmers tend to prefer agricultural machinery that is inexpensive, practical, and more mainstream [13]. Yang Xingjie’s research found that farmers with larger landholdings, less fragmented land, and higher household incomes are more willing to purchase new intelligent agricultural machinery. These farmers are more inclined to upgrade their machinery to improve crop quality and efficiency [14]. Liu Jie’s research on the policy environment found that government-led promotional campaigns and subsidies for agricultural machinery purchases have a significant impact on farmers’ adoption of intelligent agricultural machinery. Many farmers have a high level of trust in the government; when government subsidies and guarantees are available, they are highly willing to purchase intelligent agricultural machinery and participate in related training [15]. Hu Chen’s research findings conclude that farmers’ individual characteristics, cognitive traits, business environment, and policy environment all influence their adoption of new intelligent agricultural machinery and technologies [16].

Existing research has largely focused on plain areas or large-scale agricultural cooperatives, with relatively little attention paid to farmers in hilly regions. Therefore, this paper investigates the factors influencing farmers’ adoption of intelligent agricultural machinery in Changsha County, Hunan Province. From a micro-level farmer perspective, it explores the psychological perceptions and constraints regarding intelligent agricultural machinery among ordinary farmers.

Therefore, based on a field survey of farmers in Changsha County, Hunan Province, this paper uses the ordered logit model to analyze the key factors influencing the adoption of intelligent agricultural machinery by farmers in Changsha County. Based on the model results, we provided suggestions for the Changsha County government on formulating and implementing policies to promote the development of intelligent agricultural machinery.

2. Mechanism Analysis

2.1. Individual Characteristics

Farmers’ individual characteristics, such as gender, age, educational attainment, years of farming experience, and annual household income, may all influence the adoption of intelligent agricultural machinery. Generally, farmers with higher levels of education tend to possess broader knowledge and perspectives, as well as stronger cognitive abilities and a greater adoption of new technologies. In contrast, age and years of farming experience may have a negative impact. Older farmers may exhibit lower enthusiasm and ability to learn; meanwhile, longer years of farming experience imply greater production expertise, which may lead to a deeper reliance on traditional agricultural practices and, consequently, reduce acceptance of new technologies. At the same time, as the most direct economic constraint, annual household income has a positive impact on adoption behavior: the higher the annual household income, the stronger the farmer’s payment capacity and risk tolerance, and the stronger their adoption [17].

Therefore, this paper proposes the following theoretical assumptions to guide the design of the questionnaire: male farmers are more willing to adopt intelligent agricultural machinery than female farmers; older farmers are less willing to adopt intelligent agricultural machinery than younger farmers; farmers with higher levels of education are more willing to adopt intelligent agricultural machinery; farmers with longer years of farming experience are less willing to adopt intelligent agricultural machinery; and farmers with higher annual household incomes are more willing to adopt intelligent agricultural machinery.

2.2. Cognitive Characteristics

Farmers’ cognitive characteristics—that is, their subjective perceptions and evaluations of intelligent agricultural machinery technology—are key factors influencing their adoption of such technology. These characteristics primarily include the impact of intelligent agricultural machinery on agricultural production efficiency and crop yields, the difficulty of learning to operate the machinery, the purchase price, and potential risks. Specifically, when farmers gain a concrete understanding of the significant advantages of intelligent agricultural machinery in reducing costs, increasing efficiency, and improving yield and quality through on-site demonstrations, technical training, or demonstrations by neighbors, their adoption of the technology increases significantly. Conversely, if farmers perceive intelligent agricultural machinery as complex to operate, or if they view the purchase price as high and the return on investment period as long, and harbor doubts about equipment reliability and after-sales service, their adoption of such technology will be severely diminished [18].

Therefore, this paper proposes the following theoretical assumptions to guide the design of the questionnaire: farmers are more likely to adopt intelligent agricultural machinery when they believe its use helps improve agricultural productivity and crop quality and yield; their willingness decreases when they perceive it is too difficult to learn; and their adoption of it also decreases when they consider the purchase price too high or the financial risk too great.

2.3. Operating Environment

Farmers’ operational environment—that is, the objective resource conditions and market constraints they face in agricultural production—is a key factor influencing their adoption of intelligent agricultural machinery. This primarily encompasses factors such as the scale of land operations and the availability of funds for machinery purchases, which collectively determine the economic feasibility and economies of scale of intelligent agricultural machinery applications. Farmland area is one of the positive factors determining adoption. Generally speaking, the larger the farm size, the more the high purchase and operating costs of intelligent agricultural machinery can be spread out per unit area, thereby achieving a substantial return on investment. Large-scale farms or industrial enterprises thus become the mainstay of the first wave of intelligent agricultural machinery users. On the other hand, farmers’ financial capacity for purchasing machinery directly constitutes a barrier to adoption. Intelligent agricultural machinery is expensive, and even with subsidies, the actual cost remains unaffordable for ordinary farmers [19]. Therefore, farmers’ current production efficiency and profit levels are crucial. Farmers with sound business operations and higher annual household incomes not only have more ample capital but also possess a stronger risk tolerance.

Therefore, this paper proposes the following theoretical assumptions to guide the design of the questionnaire: farmers who manage larger tracts of land are more likely to adopt intelligent agricultural machinery; in regions with stronger market demand, farmers are more inclined to adopt intelligent agricultural machinery; and farmers with better financial resources are more likely to adopt intelligent agricultural machinery.

2.4. Policy Environment

The government uses fiscal incentives, services, regulations, and strategic planning to guide, incentivize, and ensure farmers’ adoption of intelligent agricultural machinery. Through systematic interventions, it effectively lowers the economic barriers, cognitive hurdles, and usage risks associated with adopting new technologies, thereby transforming potential demand into actual action. Generally, the government promotes intelligent agricultural machinery through subsidies, training, and the establishment of demonstration sites. By setting up agricultural machinery service stations to provide comprehensive services to farmers, the government enables them to benefit from government support, thereby increasing their adoption of intelligent agricultural machinery [20].

Therefore, this paper proposes the following theoretical assumptions to guide the design of the questionnaire: farmers who can benefit from more government policy incentives and related agricultural services are more willing to adopt intelligent agricultural machinery.

3. Research Design

3.1. Data Sources

Changsha County in Hunan Province was selected as the research site for this study. In terms of topography, Changsha County generally features higher elevations in the north, east, and south, with relatively gentler terrain in the central and western regions. The landscape is predominantly hilly, accounting for 89.6% of the county’s total area, while plains and mountainous areas account for 7.4% and 3.0%, respectively. Its agricultural production structure is based on grain production, with coordinated development of vegetables, rapeseed, fruits, and livestock farming. It is characterized by modernization and diversification. Furthermore, Changsha County’s gross domestic product (GDP) is projected to exceed 240 billion yuan in 2025, firmly ranking among the top county-level economies nationwide, and it has been named one of the “Top 100 Comprehensive Counties in China” for 2025. At the same time, Changsha County has ranked among China’s Top 100 New Intelligent Cities in 2025, placing within the top five nationwide. On one hand, as a national-level economic and technological development zone and Hunan Province’s first “5G + Industrial Internet” pilot zone, the county has vigorously developed the artificial intelligence (AI) industry, attracting enterprises in the AI sector and establishing innovation platforms such as the Songya Lake AI Innovation Center; on the other hand, the county has established intelligent agricultural machinery operation demonstration zones in towns such as Chunhua Town, including intelligent seedling nurseries, and has organized government-led technical training and operational demonstrations to enhance farmers’ awareness of and trust in intelligent agricultural machinery. Finally, Changsha County’s convenient transportation infrastructure facilitates large-scale household surveys, ensuring the breadth and representativeness of the sample data.

In September 2025 and January 2026, the research team conducted an in-depth and comprehensive follow-up study of the development of agricultural mechanization in Hunan Province. Through methods such as literature review, survey questionnaires, and field investigations, the team conducted questionnaire surveys in multiple towns of Changsha County (including Chunhua Town, Baisha Town, and Huangxing Town), systematically examining farmers’ adoption of intelligent agricultural machinery and the factors influencing this decision.

To determine the sample size for the survey questionnaire, we used the method described in Formula (1). Here, n is the sample size required for the survey; Z is the reliability coefficient, taken here as 1.645; p is the percentage of the population; when the true population value is unknown, it can be set to 0.5; and e is the expected margin of error, taken here as 0.06.

$$\mathrm{n}={\displaystyle \frac{{\mathrm{z}}^2\mathrm{p}(1-\mathrm{p})}{{\mathrm{e}}^2}}$$

(1)

A total of 276 questionnaires were collected during the survey. Based on the survey results, invalid questionnaires were excluded according to the following criteria: the response time should not be less than 3 min; the response options for five consecutive questions should not be identical; and answers to related questions should be consistent. Ultimately, 243 collected questionnaires were analyzed.

3.2. Variable Setup

Shao Lantong conducted research on intelligent agricultural machinery such as drones. When examining adoption behavior, they selected “willingness and unwillingness” as the variables [21]; Hao Shulei, when studying agricultural machinery cooperatives, categorized their adoption of intelligent agricultural machinery into very unwilling, somewhat unwilling, neutral, somewhat willing, and very willing [22]; Luo Jianzhang, in his research and analysis of agricultural insurance to explore farmers’ willingness to purchase, selected “unwillingness to purchase” and “willingness to purchase” as dependent variables [23]. Based on the three research methods described above, with farmers’ adoption of intelligent agricultural machinery as the dependent variable, the following scale is used: “1 = unwilling, 2 = indifferent, 3 = willing”.

When selecting independent variables, we first reviewed a large body of research by relevant scholars, including Li Rongyao’s study on factors influencing farmers’ purchase of services from agricultural cooperatives [24], and Zhou Zhou’s study on factors influencing farmers’ willingness to engage in green production practices [25], Liu Pengxin’s study on the factors influencing farmers’ adoption of variable-rate fertilization technology in Jilin Province [26], and Wang Yan’s study on the factors influencing farmers’ adoption of conservation tillage technology [27]. Based on a thorough reading and understanding of the above articles, the literature content was synthesized and summarized. Subsequently, using a combination of theoretical frameworks and field interviews, independent variables that may influence farmers’ adoption of intelligent agricultural machinery were identified from multiple dimensions, including personal characteristics, cognitive characteristics, current production and management conditions, and external environmental factors. A total of 16 independent variables were identified, as shown in

Table 1. Summary of variables.

Variable		Code	Variable Definition and Assignment
If willing to adopt intelligent agricultural machinery?		Y	1 = unwilling, 2 = indifferent, 3 = willing
Individual Characteristics	Gender	X1	1 = Male; 2 = Female
	Age	X2	1 = 0–20 years old; 2 = 21–40 years old; 3 = 41–60 years old, 4 = 61 years old and above
	Years in Agriculture	X3	1 = 1–15 years; 2 = 16–30 years; 3 = 31–45 years; 4 = 46–60 years
	Employment Status	X4	1 = Purely agricultural; 2 = Primarily agricultural; 3 = Primarily non-agricultural; 4 = Sectors other than agriculture
	Educational Attainment	X5	1 = Junior high school or below; 2 = High school/vocational school; 3 = Junior college; 4 = Bachelor’s degree; 5 = Master’s degree or higher
	Annual household income	X6	1 = 0– 50,000; 2 = 50,000–100,000; 3 = 100,000–150,000; 4 = 150,000–200,000;5 = 200,000 or more
Cognitive Characteristics	Impact on homework efficiency	X7	1 = Decreased efficiency; 2 = Almost no improvement; 3 = Some improvement; 4 = Significant improvement
	Impact on Output and Quality	X8	1 = Decreased efficiency; 2 = Almost no improvement; 3 = Some improvement; 4 = Significant improvement
	Learning Difficulty	X9	1 = Very easy; 2 = Easy; 3 = Somewhat difficult; 4 = Very difficult
	Purchase Price	X10	1 = Very low; 2 = Low; 3 = Moderate; 4 = High; 5 = Very high
	Purchase Risk	X11	1 = Very low; 2 = Low; 3 = Moderate; 4 = High; 5 = Very high
Business Environment	Farmland Area	X12	1 = 0–25 mu; 2 = 26–50 mu; 3 = 51–75 mu; 4 = 76–100 mu; 5 = 100 mu or more
	Market Demand	X13	1 = Very weak; 2 = Weak; 3 = Moderate; 4 = Strong; 5 = Very strong
	Purchase Funding Status	X14	0 = Insufficient funds, 1 = Sufficient funds
Policy Environment	Number of Government Agricultural Services	X15	1 = Very low; 2 = Low; 3 = Moderate; 4 = High; 5 = Very high
	Level of government policy support	X16	1 = Very weak; 2 = Weak; 3 = Moderate; 4 = Strong; 5 = Very strong

.

3.3. Model Construction

Farmers’ adoption of intelligent agricultural machinery is influenced by various factors. Unlike binary logit regression, ordered logit regression is suitable for situations where the dependent variable is categorical or ordinal. In this paper, farmers’ adoption of intelligent agricultural machinery is represented by “1 = unwilling, 2 = indifferent, 3 = willing.” As this is an ordered multi-class variable with a strict hierarchical order, an ordered logit model should be used for the econometric analysis. The specific formula is:

$${\mathrm{p}}_{\mathrm{i}}={\displaystyle \frac{1}{1+{\mathrm{e}}^{{-\mathrm{{\rm Y}}}_{\mathrm{i}}}}}={\displaystyle \frac{1}{1+{\mathrm{e}}^{-({\mathrm{\alpha }}_{\mathrm{i}}+{\mathrm{\beta }}_{\mathrm{i}}{\mathrm{X}}_{\mathrm{i}})}}}(\mathrm{i}=1,2,3)$$

(2)

Formula (2) can be further transformed into:

$$\ln \left({\displaystyle \frac{{\mathrm{p}}_{\mathrm{i}}}{1-{\mathrm{p}}_{\mathrm{i}}}}\right)={\mathrm{\alpha }}_{\mathrm{i}}+{\sum }_{\mathrm{j}=1}^{\mathrm{k}}{\mathrm{\beta }}_{\mathrm{i}\mathrm{j}}{\mathrm{{\rm X}}}_{\mathrm{i}\mathrm{j}}+{\mathrm{\epsilon }}_{\mathrm{i}}$$

(3)

In the formula, ${\mathrm{p}}_{\mathrm{i}}$ represents the probability that farmers’ adoption of intelligent agricultural machinery is 1, 2, or 3, respectively; ${\mathrm{{\rm X}}}_{\mathrm{i}}$ denotes the independent variables, i.e., the factors influencing farmers’ adoption of intelligent agricultural machinery; k denotes the number of independent variables; ${\mathrm{\alpha }}_{\mathrm{i}}$ is the intercept term; ${\mathrm{\beta }}_{\mathrm{i}\mathrm{j}}$ represents the coefficient to be estimated; and ε is the error term.

4. Results and Analysis

4.1. Descriptive Statistical Analysis

Descriptive statistical analyses were conducted on variables across four dimensions: individual characteristics, cognitive characteristics, the business environment, and the policy environment. The specific results are presented in

Table 2. Descriptive statistics table.

Variable Types	Variable Names	Minimum	Maximum	Average	Standard Deviation
Dependent variable Y	Whether to Adopt Intelligent Agricultural Machinery	1	3	2.765	0.513
Independent variable X	Gender	1	2	1.091	0.288
	Age	1	4	2.807	0.567
	Years in Agriculture	1	4	1.642	0.636
	Employment Status	1	4	1.391	0.629
	Educational Attainment	1	5	1.992	0.867
	Annual Household Income	1	5	3.807	1.223
	Impact on Homework Efficiency	1	4	3.370	0.612
	Impact on Output and Quality	1	4	3.288	0.623
	Learning Difficulty	1	4	2.551	0.766
	Purchase Price	1	5	4.091	0.649
	Purchase Risk	1	5	3.708	0.905
	Farmland Area	1	5	3.852	1.293
	Market Demand	1	5	2.741	0.789
	Purchase Funding Status	0	1	0.449	0.498
	Number of Government Agricultural Services	1	5	3.280	1.245
	Level of government policy support	1	5	3.317	1.092

.

4.2. Model Reliability Testing

We used IBM SPSS Statistics 27 software to test for common method bias in the questionnaire data using Harman single-factor test. And the SPSS factor analysis shows that the variance of the first factor is 25.157%, which is less than the critical value of 40%, and thus, there is no significant common method bias in this result.

We then conducted a multicollinearity test on the model. Generally, the larger the variance inflation factor (VIF), the greater the multicollinearity among the variables; conversely, the smaller the VIF, the less severe the multicollinearity problem. Generally, if the VIF is greater than 10, it indicates a significant multicollinearity problem among the variables; a VIF value between 5 and 10 suggests a certain degree of correlation among the variables; and a VIF value less than 5 indicates weak multicollinearity among the variables. Based on the results of the multicollinearity test in

Table 3. Results of the multicollinearity test.

Variable Name	VIF
Gender	1.228
Age	1.735
Years of farming	1.446
Employment status	1.397
Educational attainment	1.749
Annual household income	2.147
Impact on work efficiency	3.293
Impact on yield and quality	3.538
Learning difficulty	2.326
Purchase price	2.044
Purchase Risk	1.804
Farmland area	3.009
Level of market demand	2.107
Purchase funding status	1.369
Number of government agricultural service visits	3.184
Level of government policy support	2.326

, the maximum VIF value among the 16 variables included in the study is 3.538, which is less than 5. This indicates that there is no multicollinearity issue among the variables in this study’s model, and logistic regression analysis can be applied directly.

As shown in

Table 4. Model fitting information.

Model	−2 Log-Likelihood	Chi-Square	Degrees of Freedom	Sig	AIC	BIC
Intercept only	287.347
Final	110.623	176.725	19	<0.001	151.626	214.501

, the results of the model fit tests show that the −2 Log Likelihood value of the Final Model is significantly lower than that of the model containing only the intercept term. The chi-square value is 176.725, and the chi-square test is significant (p < 0.05). The AIC and BIC values are used to compare results across multiple analyses; therefore, the lower these values are, the better. This indicates that the ordered logit model developed has a good overall fit, and the explanatory variables effectively account for variations in farmers’ willingness to adopt intelligent agricultural machinery.

As shown in

Table 5. Pseudo R² Inspection.

Pseudo R2	Value
Cox and Snell R2	0.517
Nagelkerke R2	0.745
McFadden R2	0.615

, the three fit statistics for this research model are as follows: McFadden R² = 0.615, indicating that the 16 selected independent variables account for 61.5% of the variation in farmers’ adoption of intelligent agricultural machinery; Cox and Snell R² = 0.517, and Nagelkerke R² = 0.745, suggesting that the model fits the data quite well.

As shown in

Table 6. Test for parallel lines.

Model	−2 Log-Likelihood	Chi-Square	Degrees of Freedom	Sig
Null Hypothesis	110.623	-	-	-
General	85.992	24.631	19	0.173

, the significance value for the parallel lines test in this model is 0.173 (>0.05); therefore, the model holds, and the ordered logit model can be used for analysis.

4.3. Analysis of Ordinal Logit Model Results

The analysis results of the ordered logit model are shown in

Table 7. Results of the ordered logit analysis.

	Coeff	Std Err	Wald	p-Value
Willingness to use smart agricultural machinery = 1	−31.682	6.674	22.531	0.000
Willingness to use smart agricultural machinery = 2	−27.775	6.589	17.767	0.000
Gender (male)	−3.912	1.456	7.220	0.007
Gender (female)	0 a	-	-	-
Age	0.770	0.579	1.765	0.184
Years of farming	−0.505	0.513	0.971	0.324
Employment status (Purely agricultural)	−5.505	1.779	9.580	0.002
Employment status (Primarily agricultural)	−5.116	2.075	6.076	0.014
Employment status (Primarily non-agricultural)	0 a	-	-	-
Educational attainment (Junior high school or below)	−20.557	1.292	253.284	0.000
Educational attainment (High school/vocational school)	−20.054	1.325	229.204	0.000
Educational attainment (Bachelor’s degree)	0 a	-	-	-
Annual household income	−0.859	0.553	2.412	0.120
Impact on work efficiency	1.689	0.969	3.041	0.081
Impact on yield and quality	−0.864	1.063	0.660	0.416
Learning difficulty	−0.478	0.709	0.454	0.500
Purchase price	0.602	0.628	0.919	0.338
Purchase Risk	−1.726	0.620	7.741	0.005
Farmland area	−0.843	0.565	2.225	0.136
Level of market demand	1.926	0.535	12.950	0.000
Purchase funding status (Insufficient funds)	−0.102	0.794	0.017	0.898
Purchase funding status (Sufficient funds)	0 a	-	-	-
Number of government agricultural service visits	2.296	0.674	11.618	0.000
Level of government policy support	−0.012	0.621	0.000	0.985

a: This parameter is redundant, so set it to 0.

. The impact of gender on farmers’ adoption of intelligent agricultural machinery. According to the regression results, the p-value for male farmers is 0.007, and the regression coefficient is −3.912, indicating that the gender variable has a significant impact on farmers’ adoption of intelligent agricultural machinery. Taking female farmers as the reference group, the coefficient for male farmers is negative, suggesting that male farmers have a lower willingness to adopt intelligent agricultural machinery than female farmers. According to the survey, in recent years, the outflow of the rural labor force has intensified, with a large number of male laborers turning to non-agricultural employment, and women gradually becoming important participants in agricultural production. Therefore, women have a more obvious demand for improving agricultural production efficiency and reducing labor pressure, and thus female farmers have a higher acceptance of intelligent agricultural machinery.

The impact of occupation status on farmers’ adoption of intelligent agricultural machinery. According to the regression results, taking “non-agriculture as the main occupation” as the reference group, both “pure agriculture” and “agriculture as the main occupation” farmers passed the significance test. Among them, the p-value for pure agriculture farmers is 0.002, and the regression coefficient is −5.505; the p-value for farmers mainly engaged in agriculture is 0.014, and the regression coefficient is −5.116. This indicates that the possibility of these two types of farmers adopting intelligent agricultural machinery is lower than that of farmers mainly engaged in non-agriculture. The survey found that pure agriculture farmers have a relatively single source of income, and the risk of agricultural production is relatively high. They often hold a cautious attitude towards high-investment equipment. At the same time, they have formed a strong inertia towards traditional agricultural production methods, leading to a certain degree of wait-and-see mentality towards intelligent agricultural machinery.

The impact of educational level on farmers’ adoption of intelligent agricultural machinery. According to the regression results, taking farmers with a bachelor’s degree as the reference group, the p-value for farmers with junior high school education or below is 0.000, and the regression coefficient is −20.557; the p-value for farmers with a high school education is 0.000, and the regression coefficient is −20.054. This indicates that the educational level variable has a significant impact on farmers’ adoption of intelligent agricultural machinery. From the survey, farmers with a higher educational level have obvious advantages in obtaining and understanding new technology information, and their acceptance of new agricultural technologies and intelligent equipment is also relatively stronger. In the agricultural production process, such farmers are often more likely to learn about the functions and application methods of intelligent agricultural machinery through training, online platforms, or agricultural technology promotion activities, thereby enhancing their recognition of new technologies and being more willing to adopt intelligent agricultural machinery.

The impact of purchase risk on farmers’ adoption of intelligent agricultural machinery. According to the regression results, the p-value for purchase risk is 0.005, and the regression coefficient is −1.726. This indicates that the purchase risk variable has a significant negative impact on farmers’ adoption of intelligent agricultural machinery. Intelligent agricultural machinery is a high-investment agricultural piece of equipment, and its purchase cost is significantly higher than that of traditional agricultural machinery. For most farmers, purchasing it entails a significant financial burden. At the same time, the technology update speed of intelligent agricultural machinery is relatively fast, and farmers are worried that the equipment will become technologically outdated in a short period of time, thereby affecting the value of the equipment. In addition, the maintenance cost of intelligent agricultural machinery is relatively high, and in some areas, there is an incomplete after-sales service system. Farmers are worried that equipment failures will be difficult to repair and the repair costs will be high, thereby increasing risk perception and reducing the willingness to purchase.

The impact of market demand on farmers’ adoption of intelligent agricultural machinery. According to the regression results, the p-value for market demand is 0.000, and the regression coefficient is 1.926. This indicates that the higher the market demand for intelligent agricultural production, the stronger the willingness of farmers to adopt intelligent agricultural machinery. As consumers’ demand for green and high-quality agricultural products keeps increasing, agricultural production is gradually moving towards standardization and precision. Smart agricultural machinery can effectively enhance the precision of agricultural production and thus better meet the demands of modern agriculture. Therefore, with the continuous growth of market demand, the smart agricultural machinery industry will develop rapidly, and the possibility of farmers adopting smart agricultural machinery will increase significantly.

The impact of agricultural services provided by the government on farmers’ adoption of smart agricultural machinery. According to the regression results, the p-value of agricultural services is 0.000, and the regression coefficient is 2.296, indicating that the more frequently the government provides agricultural technical services, the more likely farmers are to adopt smart agricultural machinery. The government’s organization of agricultural technology training, on-site demonstrations, and promotional activities can enhance farmers’ understanding of smart agricultural machinery and reduce their unfamiliarity with new technologies. At the same time, an increase in the frequency of agricultural services also implies a more complete grassroots agricultural extension system and smoother information communication between the government and farmers, which is conducive to enhancing farmers’ trust in smart agricultural machinery and increasing the possibility of their adoption.

5. Conclusions and Policy Recommendations

5.1. Research Findings

Male farmers, those engaged purely in agriculture or mainly in agriculture, and those with a high school education or below, are less likely to adopt intelligent agricultural machinery. Meanwhile, as the risk of purchasing intelligent agricultural machinery decreases, market demand increases, and the number of agricultural services provided by the government increases, the possibility of farmers adopting intelligent agricultural machinery also increases.

5.2. Policy Recommendations for the Changsha County Government

(1) Farmers with higher educational attainment exhibit a stronger willingness to adopt intelligent agricultural machinery. Thus, enhancing farmers’ digital literacy and equipment operation proficiency constitutes a fundamental prerequisite for advancing the popularization of intelligent agricultural machinery. For farmers with lower educational levels, training should prioritize basic operational skills, employing an approach of “on-site demonstration plus hands-on instruction” to minimize theoretical exposition and emphasize practical training. Concurrently, townships may select a cohort of farmers with higher educational backgrounds and stronger learning receptivity as demonstration users of intelligent agricultural machinery. Following standardized training by agricultural and rural affairs departments, these demonstrators will assume responsibility for providing technical guidance to neighboring farmers. Training venues should be strategically located in township agricultural machinery service stations, cooperatives, or field sites to avoid long-distance travel for farmers and thereby improve participation rates.

(2) Farmers engaged in pure agriculture or agriculture-dominated livelihoods should be prioritized for inclusion in the agricultural machinery purchase subsidy list. This entails increasing the subsidy ratio for intelligent agricultural machinery while appropriately reducing the proportion of self-raised funds required. Additionally, the “shared agricultural machinery” model should be promoted: agricultural machinery cooperatives are encouraged to offer shared services for intelligent agricultural machinery, with charging mechanisms based on acreage or time-based leasing. This approach reduces farmers’ one-time purchase costs and lowers the threshold for adoption.

(3) Farmer’ willingness to adopt intelligent agricultural machinery is inversely correlated with their perceived purchase risks. To mitigate this, the government may establish specialized insurance schemes for intelligent agricultural machinery, including equipment failure insurance, operational loss insurance, and natural disaster loss insurance. Furthermore, the after-sales service system should be optimized by establishing multiple maintenance outlets and spare parts supply centers within counties, thereby alleviating farmers’ concerns about post-purchase support.

(4) Higher market demand for smart agriculture corresponds to greater farmer willingness to adopt intelligent agricultural machinery. Accordingly, a core demonstration zone for smart agriculture or an application demonstration base for intelligent agricultural machinery can be established in Changsha County. This base will centrally showcase unmanned agricultural machinery, intelligent monitoring systems, and precision operation technologies, with real-world benefit cases used to enhance farmers’ confidence in the market prospects of intelligent agricultural machinery.

(5) The frequency of government agricultural services is positively associated with farmers’ adoption willingness. To strengthen this, the staffing quota for grassroots agricultural machinery technicians in Changsha County should be expanded, and regular specialized training on intelligent agricultural machinery should be conducted to improve the digital agriculture capabilities of grassroots extension personnel. Simultaneously, online platforms for intelligent agricultural machinery consultation and fault reporting can be developed via WeChat, mini-programs, and other digital tools, enabling farmers to access timely assistance.

5.3. Limitation of the Study

The primary research site was Changsha County; although somewhat representative, the findings may not be applicable to remote areas of Hunan Province. Future research could expand the study area to cover the entire province of Hunan, increase the sample size, and enhance the study’s reliability.