Switched reluctance motor (SRM) and its control system, after decades of development, has been better established. Its inherent merits such as simple and sturdy structure and high reliability could contribute to its excellent running performance and high performance/cost ratio for electric vehicle application. A U-type modular double-stator SRM (DSSRM), which has the merit of high power density is optimized in this paper. Taking average torque, torque smoothing coefficient and efficiency as objectives, sensitivity analysis on DSSRM is investigated and discussed to select significant geometric variables. On the basis of the initial structural parameters, the multi objective optimization design is processed by adopting a fuzzy iteration optimization algorithm with weight determination. After the determination of decision result of each structure parameter, the best dimension scheme of the DSSRM is concluded. Moreover, the final scheme proves to perform better than the initial scheme by finite element (FE) analysis. Finally, a prototype motor is manufactured and the experimental results validate the improvement of the DSSRM performance.