This paper covers driver's control algorithm changing from forward vehicle following to lane change situation. Accelerator control was classified by control target as active control, passive control and neutral. It was clear that active control was most participating process for control target. Next, control algorithm changing was expressed by relationship between even control process and relative velocity. Therefore, using a neural network system, driver's control behavior is simulated by constructing a driver model. And it was found that this model could accurately reproduce driver's control behavior of even control algorithm. After that, driver's control characteristics ware showed by factor analysis method that was used to calculate the contributing rate of inputs for driver's control behavior.