The fleet management service identifies the driver behavior pattern, restricts improper operation and avoids trouble-making actions or deviant behavior to save operation costs. The fleet management system architecture mostly depends on hardware driving recorders to obtain relevant data. However, the hardware system is not easy to customize, expand or adjust due to its fixed data categories and system functions. Therefore, by integrating OBD-II and smart phones, this paper proposes developing a software vehicle tracking system with an algorithm to detect vehicle-environment interactive behaviors based on Android, an open source software stack. This system aims to collect various information from vehicles and mobile phone sensors, identify interactive behavior, provide the back-end vehicle management system for monitoring and analysis, improve traditional fleet management in terms of customization, expansion and real-time performance, and facilitate fast-speed establishment and customization of the driving management system.