Establishing an accurate switched reluctance motor (SRM) model has an important impact on improving the performance and control effect of SRM. For the problems of high saturation and serious nonlinearity of magnetic circuit during SRM operation, an SRM nonlinear model based on back propagation (BP) neural network algorithm optimized by mind evolutionary algorithm (MEA) is proposed. A four-phase 8/6-pole SRM model is established by using ANSYS Maxwell software and the finite element calculation is carried out. It is verified through the comparison between the simulated value and the experimental value that the model has higher accuracy than the BP neural network model without MEA optimization. It can better reflect the flux linkage and torque characteristics during SRM operation, and has better generalization ability.