In order to improve the working performance of the magnetorheological damper (MRD), this paper has carried out the optimization design of the MRD structure. Firstly, a preliminary design of the MRD magnetic circuit structure was carried out with the output damping force as the design objective, and the magnetic field analysis of the magnetic circuit structure was carried out by finite element simulation software. The magnetic flux densities of each part under different combinations of magnetic circuit structure parameters were obtained through several simulation tests. Based on the finite element simulation results, a nonlinear mapping model between the magnetic circuit structure parameters and the magnetic flux density was constructed using a back propagation (BP) neural network. Then, taking the maximum adjustable multiplier of the damping force of the MRD as the objective, and the maximum damping force and the flux densities of each part as the optimization constraints, the established magnetic field neural network model was combined with the whale optimization algorithm to realize the intelligent optimization of the MRD structure. Finally, the performance test of the optimized MRD was carried out by building an experimental platform, and the experimental results showed that the measured performance of the damper matched well with the theoretical calculation results, which effectively verified the effectiveness and practicality of the proposed optimization design method.