Abstract—In this paper, we designed a android application to collect real-time vehicle data and analyze the driving habits. Vehicle data, such as speed, and acceleration can be collected from an On-Board Diagnostics module and transferred to a mobile phone through Bluetooth, and then deliver the data from the mobile phone to the backend computing server using HTTP over GPRS or 3G/4G mobile communication networks. An Error Back-Propagation Algorithm (BP neural network algorithm) is developed to classify and analysis the vehicle data to achieve the evaluation on driving habits.