Understanding human mobility patterns is key to developing intelligent, context-aware services. This study analyzes GPS data from users’ daily activities to identify personal POIs, frequently visited places such as home, workplace, or leisure locations. Using spatial-temporal clustering, we detect these POIs and model users’ mobility behavior. We then apply predictive modeling techniques to forecast the next likely POI a user will visit. The objective is to enable proactive, data-driven mobility assistance that supports personalized and efficient daily movement planning. The dataset includes GPS trajectories from 288 users collected over three months (March, April, and May), and the proposed model achieves a prediction accuracy of 90%.