In recent years, the Beijing Municipal Government has implemented the Paddy Land-to-Dry Land (PLDL) program in the Miyun Reservoir catchment, in order to ensure the security of Beijing's water supply, and the participation of rural households is the key to the program's success. Based on survey data of rural households in three townships of Fengning County (Hebei Province), the present study used choice experiment modeling to quantitatively estimate the effects of six choice attributes, including contract length, contract release option, proportion of total land area enrolled, mandatory fertilizer use reduction, allowance of irrigation, and subsidy level, on rural households' willingness to participate in the PLDL program. Except for the proportion of total land area enrolled, all the choice attributes significantly affected the rural households' willingness to participate. Based on the initial result, the marginal effect and households' marginal willingness to accept for each of the significant choice attributes was estimated. The results indicated that (1) rural households preferred the option to be released from their contracts and shorter contract length, thereby identifying a sustainability risk for the program, which could be addressed by creating more off-farm employment opportunities in accordance with the characteristics of local labor forces. (2) The restriction of irrigation and mandatory fertilizer use reduction both inhibit program participation, and in the current phase, these issues should be addressed by improving rural households' environment awareness through relevant technical training to improve their production activities. (3) Meanwhile, subsidy level limited the motivation of rural household participation, whereas the other choice attributes were more effective, which suggests that allowing for greater household autonomy in program participation and alternative contractual arrangements should be piloted. (4) Finally, the willingness to participate in the PLDL program varied across different townships, and the township with non-farm industries had a higher compensation expectation than the agricultural township; therefore, the current one-size-fits-all compensation strategy should be replaced with one that considers regional economic characteristics.