As a new model supersedes the conventional wave forecast methods, Multiple Regression wave forecast models described in PHysical parameters (MRPH model) is developed. In the MRPH model, ocean waves are separated into wind waves and swells, and assumed that the propagation velocity of each wave components is constant respectively.
　The MRPH model shows more excellent accuracy of forecasting, and solves the difficulties that the presence of lag time which is defects of conventional multiple
regression model. Also, accuracies of the MRPH model does not decreases remarkably in the case of long term wave forecasting. The MRPH model is able to predict wave period, direction, and the components of wind waves and swells which have not predicted by usual statistical model.