This paper presents an improved characteristic analysis methodology for a 5-phase hybrid stepping motor. The basic approach is based on the use of equivalent magnetic circuit taking into account the localized saturation throughout the hybrid stepping motor. The finite element method(FEM) is used to generate the magnetic circuit parameters for the complex stator and rotor teeth and airgap considering the saturation effects in tooth and poles. In addition, the neural network is used to map a change of parameters and predicts their approximation. Therefore, the proposed method efficiently improves the accuracy of analysis by using the parameter characterizing localized saturation effects and reduces the computational time by using the neural network. An improved circuit model of 5-phase hybrid stepping motor is presented and its application is provided to demonstrate the effectiveness of the proposed method.