Model predictive torque control for permanent magnet synchronous motors is sensitive to parameters and requires weighting factors. Model-free parallel predictive torque control (MF-PPTC) is one of the latest solutions. However, it suffers from a heavy computational burden and hardly achieves the optimization of both torque and flux. In this article, a parameter-free predictive torque and flux control (PF-PTFC) is proposed to improve parameter robustness and eliminate weighting factors. First, a parameter-free incremental stator flux predictive model (ISFPM) is defined by eliminating parameter terms found in conventional predictive models. Based on ISFPM, a new cost function without weighting factors is constructed to automatically compensate for the impact of missing parameter items. Compared with the MF-PPTC, the proposed strategy has a lower computational burden and better steady-state performance. Moreover, the PF-PTFC is combined with an active-disturbance-rejection-based discrete-time integral sliding-mode speed controller to improve speed regulation. The experiments confirm the proposed method's superiority.