The frequency of distracted driving and its impact on safety is rapidly become a serious social issue. Given that distracted drivers pose an increased crash risk not only to themselves, but also to other road users, it is important to investigate ways to address this growing issue. This research project aims to realize a real-time identification and detection system for distracted driving for use in real-world driving scenarios. Therefore, the main goal of this research is development of a real-time model using on-board sensors, to classify whether the driver of the leading vehicle is distracted or not. This research also investigates the typical types of distracted driving behavior that can be detected by host vehicles (depending on the available sensors) and their key characteristics on driving data pattern.