With recent emerging events such as the pandemic of coronavirus disease (COVID) 2019, human mobility has caused significant concern in the spread of this dangerous pandemic, so mobility prediction is considered as one of the crucial factors to prevent the pandemic. Therefore, there have been many proposed and highly functional studies. Applications of social networks have stored vast data of user movements and brought a vast of interesting research on human mobility. Friendship on social networks has also revealed some effects on the movement. In this study, we have attempted to explore the influence of friendships in location-based social networks on human mobility. We conduct the movement based on the K latest check-ins of friends of the user to predict mobility. We have deployed Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm (using Haversine distance) to cluster original check-in points and filtered the K latest friends’ check-ins of the user to predict the user’s next movement with the Random Forest algorithm. The prediction conducted from movement history of friends has obtained better performances compared to the prediction without considering the Friendship. The highest accuracy is 0.3176 (with a radius of 400 m and four latest check-ins of friends). Besides, we compare and evaluate the results of the proposed method with the clustered dataset with the original dataset. As observed from the experiments, clusters generated by DBSCAN with wider radii can reveal that their friends’ movements can influence users’ mobility on a location-based social network (LBSN).