Traffic simulation is a useful tool to study human factors, traffic system, and vehicle design. To build a realistic and intelligent traffic environment, driver behavior models need to be studied. A significant problem in modeling driver behavior was the lack of real traffic data. This thesis presents an experiment for obtaining real traffic data and shows the filtering process for the data. This is the first experiment using a rear-looking radar-tracking device to obtain real traffic data. Based on the data, some driver behavior models, such as car-following models and lane-changing models, have been constructed and some simulation results have been presented and compared. Furthermore, the traffic simulation with driver behavior models has been implemented into the driver-simulator system at Iowa State University. Current developments of the driver-simulator system have been discussed.