Recently optimal regulator theory has been applied to the control of the manoeuvring motion of a ship. But it needs accurate mathematical model of the controlled object and it has not sufficient controllability to the nonlinearity of the controlled object. By these reason it was tried by one of authors to apply the Learning Feed-Forward Control (LFFC) system to the follow-up control to the desired value for the ship manoeuvring motion. The LFFC system is a kind of the neural network model. It is not a multi-layered perceptron type but a kind of an adaptive filter, and it has a dynamic quality. The system is tuned with the feedback-error-learning method proposed by Kawato and others. It was recognized that the LFFC system had a good controllability and the problem in the optimal regulator system mentioned above was solved. The servo mechanism needs both an ability of the follow-up control to the desired value and that of the compensation of the influence from the disturbance. So in this paper it is tried to apply the LFFC system to the compensation of the influence from the disturbance to the ship manoeuvring motion. And for the basic study the following case is investigated with the computer simulation. That is, the heading angle of the ship is controlled with the bow thruster in the wind disturbance. It becomes clear that the system has a good controllability to the compensation of the influence from the disturbance due to the self-tuning ability and a feed-forward loop.