[Objective] To solve the problem of poor robustness in the model predictive torque control (MPTC) algorithm caused by parameter mismatch in the control model of permanent magnet synchronous motors, an MPTC algorithm based on an unweighted factor proportional integral derivative (PID) cost function was proposed. [Methods] The PID cost function eliminated static error by constructing an integral error value function and suppressed the oscillation of torque and flux linkage errors by constructing a differential error value function. However, since torque and flux linkage had different dimensions, the PID cost function still included weight coefficients. To solve this problem, an unweighted dual-cost function parallel strategy was proposed, which transformed the multi-objective cost function into a single-objective cost function to obtain a set of voltage vectors. The current-type cost function was used as the final criterion to evaluate the optimal voltage vector, thus eliminating the need for weight coefficient. Finally, the proposed unweighted factor PID cost function MPTC was compared with the traditional cost function MPTC through simulations. [Results] The results show that MPTC with unweighted factor PID cost function can suppress torque and flux ripple, reduce the dependence of MPTC on model parameters, and solve the difficulty in adapting to different operating conditions because of the existence of weight coefficient, while maintaining the advantage of MPTC’s fast dynamic response. [Conclusion] The proposed cost function demonstrates good feasibility in MPTC for permanent magnet synchronous motors.