Abstract: How to promote the rapid development of the agricultural economy while taking into account environmental protection and building a "community of life" between humans and nature has become an important issue that we can't avoid in promoting China's rural vitalization strategy. Different from existing studies, in this paper, we used the propensity score matching (PSM) estimation technique to overcome the endogenous nature of sample selective bias, and the "strong embeddedness" factor in sociology. As such, the impact of relational embeddedness and structural embeddedness on the green disposal behavior of farmers' agricultural materials wastes were explored. The result showed that peasant household concurrent business behavior and per capita income of farmers' household all played a significant positive role on affecting relational embeddedness. That was, the relationship strength and the quality of the relationship were more positively affected by the peasant household concurrent business behavior and the per capita income of the family. At the same time, the number of family labor, peasant household concurrent business behavior, the existing cultivated land area of the family, and the proportion of agricultural income would significantly affect the structural embeddedness. Then the K-nearest neighbor matching method was used to explore the difference between the green disposal behavior of the agricultural materials waste of the control group and the treated group. According to the propensity scores of the control group and the treated group, it was clear that the probability of occurrence of green disposal of agricultural materials waste in the farmers' households in the control group was 4.10% before the matching; after matching, the proportion increased to 6.60%. On the whole, the relational embeddedness would increase the probability of the farmers' agricultural materials waste green disposal behavior by 4.70%. Therefore, based on the survey data of this paper, we did not find that the relational embeddedness had a significant impact on the green disposal behavior of agricultural household materials waste. Similarly, K-nearest neighbor matching method was also used to explore the relationship between structural embeddedness and the green disposal behavior of agricultural household materials waste. After matching, the occurrence probability of the green disposal behavior of agricultural household materials waste increased by 10.10%. At the same time, the t-statistic of average effect of treatment on the treated (ATT) was greater than 1.96 (P<0.05). This showed that there was a significant difference in the green disposal behavior of agricultural materials wastes between farmers who were in the state of structural embeddedness and those who were not in the state of structural embeddedness. In addition, in order to ensure the robustness of the research results, the "self-help method" (Bootstrap) was used in the matching process to infer the overall standard error. According to the results, average effect of treatment on the treated (ATT), average effect of treatment on the untreated (ATU) and average effect of treatment on the population (ATE) were basically consistent with the above results, whether in the state of relational embeddedness or structural embeddedness. And the results of other matching methods, including K-nearest neighbor matching method, radius matching method, kernel matching method and Mahalanobis matching method, were highly consistent with the above results, which also can identify the strong robustness of the research results. Based on this, we suggested that through the establishment of peasants' interests appeal channel and accountability supervision mechanism, farmers should be encouraged to strengthen exchanges and communication with professional farmers, and create a good social atmosphere of environmental protection to improve the green disposal behavior of farmers' agricultural materials waste, and promote the green transformation of the agricultural industry.