The coronavirus disease of the 2019 (COVID-19) pandemic has increasingly spread worldwide with tremendous damage. The human movement can make this infectious disease more contagious and become a primary concern for controlling the spread of COVID-19. Therefore, human mobility prediction, which holds an essential role in numerous applications (e.g., estimating migratory flows, traffic forecasting, urban planning, etc.), is now even more urgent for preventing the pandemic. This work presents a human mobility prediction approach based on movement patterns with k-Latest Check-ins (kLC) and an evaluation of the different radius to cover related areas for the prediction. The proposed method is evaluated on more than six million human move- ment history records of more than 70,000 users checking in at more than 168,000 locations and achieved promising results compared to the state- of-the-art. The results reveal that most of the users in Brightkite move around within about 20 familiar places with a mobility prediction accu- racy reaching 90%. In contrast, Gowalla’s users tend to extend their movement with further distances.