Based on the static mathematical model of Preisach hysteresis, this paper presents a modified method, combining Genetic Algorithm and Neural Network Method to get three conventional parameters of the Preisach hysteresis model under alternating magnetization. In the proposed method, considering the symmetry of the major hysteresis loop, a complete fitting hysteresis loop is obtained through the neural network training to find the optimization function and through the genetic algorithm to find the function extreme. The numerical fitting results obtained by this method are found to be in good agreement with the measurement data and verify the theory.