 Our study examined the mobility patterns of over 10 million smart device users in three major Chinese cities during the COVID-19 pandemic. We found that the mobility patterns were remarkably similar across the three cities despite differences in geography and epidemic intensity. Our proposed model was able to accurately predict human mobility by incorporating COVID-19 statistics and policies. The model can be used to help policymakers understand the impact of different policies on human mobility and mitigate the pandemic. This article was authored by Li Qihong, Li Chao, Xiang Yinfeng, and others.