Abstract: The fast development of positioning technology and mobile computing is changing the way people live. People can connect the internet anytime and anywhere and it brings us amounts of GPS data representing the users’ location histories. A system was designed to mine periodic behaviors on GPS data. In the stage of data collection, smartphones running on Android will be used to collect GPS data and then the data will be sent to a HTTP server at regular intervals. In the data mining stage, it is aimed to mine reference spot and understand the periodic pattern based on different users’ GPS data. Also, stay points were used to represent the behaviors of users. With the sequence of stay points, frequent periodic patterns were mined and then these patterns were summarized. Finally, the experiments indicate our system shows efficient and stable performance with large numbers of GPS data.