Recently many kinds of four-wheel-steering (4WS) systems have been studied and proposed. Design of these control systems is based mainly on the two-degrees-of-freedom model. However, vehicle motion is influenced by various nonlinear factors, for example, tire characteristics and road condition. Hence, it is difficult to express the vehicle motion accurately by the two-degrees-of-freedom model. In this paper, a method of 4WS control system is presented by applying a neural network. First we identify the vehicle motion as a system neural network. Next we construct a neural network controller by using this system neural network. The effectiveness of the proposed method is demonstrated by a computer simulation.