This research deals with a construction of driver model to express driving behavior at obstacle avoidance. A risk potential model is constructed with a control algorithm including a target course determination process. The steering angle of feedforward control is calculated using curvature of risk minimum point at the risk potential map corresponding to the target course. On the other hand, the steering angle of feedback control is calculated using the difference between left- and right-hand sides of the risk potential at prediction point. As a result, the risk potential model can describe the driver behavior when obstacle avoidance.