In-depth study of the driving behavior and characteristics of drivers is of great significance in promoting the development of intelligent vehicle to highly automatic drive. The application of driving simulator to the study of the driver's driving decision and driving behavior in complex traffic scenarios has become a research hotspot in recent years. Based on the unreal engine 4(UE4) interactive scene simulation technology, we built a driving simulation environment with many traffic elements such as vehicles, roads, buildings, traffic lights, pedestrians, road signs, etc. and designed the high-fidelity and human-machine interactive driving simulation system of Logitech G29 force feedback steering wheel pedal set with CarSim vehicle dynamics model. With the driving simulation system, driving simulation experiments were conducted under typical traffic scenarios and driver's information and real-time driving data were collected to study driving behavior and characteristics. The results show that by the driving simulator a realistic driving experience can be got. Based on Fuzzy C-means clustering algorithm (FCM), the driver characteristics are classified accurately into six clusters, the driver behavior and driving style can be analyzed, and thus the association between driver characteristics and driving capacity are established. An effective way is provided for further research on the allocation of driving rights concerning real-time driving ability.