In order to avoid the design and tuning of the weighting factors in conventional cost function, the ranking approach is used to implement the model predictive torque control (MPTC) for surface permanent magnet synchronous motor (SPMSM). The control variables with different dimension are converted to dimensionless ranking scores, therefore, the weighting factors are eliminated. The priority rules are given to solve the problem that the optimal voltage vector is not unique for ranking approach, and the effects of different priorities on control performances are analyzed. Simulation results verify the effectiveness of the ranking approach in the SPMSM MPTC system.