Abstract: Influencing factors of farmers’ willingness on homestead replacement (FWHR) were usually detected based on single level in previous studies, in which multi-level effects of the factors were ignored. To solve this issue, questionnaire survey was conducted on 394 households from 17 villages in Suzhou and Changzhou, which are located in the developed area, and multi-level factors’ effects on FWHR were analyzed by using multi-level regression model. Results show that the farmers’ willingness in Suzhou is lower than that in Changzhou, and significant difference of FWHR exists in different villages, 75.12% of which can be attributed to the differences among villages. The factors of age, education level, and the main source of household’s income have significant positive effects on FWHR, while the family size has negative effect on FWHR. The village’s average income of households and the environmental level significantly weaken the positive relationship between farmers’ main income source and FWHR, while the level of village’s economic development and distance to the nearest city or town strengthen the aforementioned positive relationship. In addition, the factors from the village-level explain 95.69% of the correlation between farmers’ main income source and FWHR. It is indicated that factors from village-level lead to the difference of the FWHR among villages through changing the influence degree of the factors from farmer-level. Therefore, to promote the work of farmers’ homestead replacement, the influencing factors from farmer-level and village-level should be considered comprehensively, the relevant supporting policies and social security system should be improved. Meanwhile, the publicity of rural homestead replacement polices should be strengthened, the rural homestead planning should be optimized. In addition, the public service facilities and the environment in rural area should be improved, as well as the farmers’ income should be increased.
Key words：rural homestead; replacement willingness; multi-level logistic regression model; village level; farmer level