Aiming at the design and adjustment of the weighting factors for model predictive torque control (MPTC) of permanent magnet synchronous motor (PMSM), the parallel form of multiobject cost functions connected by weighting factors is converted into the cascaded form of singleobject cost function, therefore to eliminate the weighting factors. The MPTC for surface PMSM based on cascaded method is established to study the effect of cascaded sequence and the number of output voltage vectors on the system control performance and calculation quantity. The method to simply the sorting comparison calculation and the optimization strategy to dynamically adjust the number of output voltage vectors using a fuzzy controller are proposed. The simulation results show the MPTC for surface PMSM based on cascaded method is feasible without the design and adjustment of weighting factors. When the first stage outputs 6 voltage vectors, the number of sorting comparison and the running time of the first stage using the simplified algorithm are reduced by 71.43% and 71.67% respectively, and the total running time is reduced by 2.28%. Compared with conventional cascaded method, the fuzzy dynamic cascaded MPTC can dynamically adjust the importance of different control objectives, decrease torque and stator flux ripple, suppress dynamic flux ripple, and optimize control performance.