Rural-to-urban migration has increased rapidly in China since the early 1980s, with the number of migrants has reached 376 million by 2020. Despite this sharp trend and the significant contributions that migrants have made to urban development, the migrant workers have had very limited access to the social insurance that the majority of urban workers enjoy. Against the background of the social insurance system adjustment in Chengdu in 2011, this study uses a difference-in-differences (DID) model to empirically test the impacts of changes in the social insurance policy contribution rates on the social insurance participation rates of migrant workers, using the China Migrants Dynamic Survey (CMDS) data for 2009–2016. We find that the social insurance participation rate of migrant workers was significantly reduced after they were incorporated into the urban worker insurance system. There was no significant change in the wages of migrant workers, but the working hours were increased and their consumption level decreased. In other words, simply changing the social insurance model of migrant workers from “comprehensive social insurance” to “urban employee insurance” reduces the incentives for migrant workers to participate in insurance and harms the overall welfare of migrant workers. Our study indicates that the design of the social security policy is an important reason for the lower participation rate of migrants. It is necessary to solve the problem of insufficient incentives through the targeted social security policies; primarily, the formulation of a social security policy contribution rate suitable for the migrants, and the establishment of a comprehensive social security policy and the gradual integration of the social security system.