[Objective] To address the heavy computational burden of traversing all voltage vectors in two-step model predictive current control (MPCC) for permanent magnet synchronous motor (PMSM), this study proposes three streamlined control sets. These sets reduce the number of voltage vector traversal by incorporating additional constraints, thereby improving real-time performance. [Methods] Firstly, the voltage vectors were classified into five categories based on the positive and negative signs of their projections on the dq-axis. Subsequently, the error between the reference and actual values of the motor stator dq-axis current was used as constraints. The selection patterns of these five voltage categories were analyzed for both the first and second steps of model prediction under the imposed constraints. By eliminating underutilized voltage vectors, streamlined control set 1 and streamlined control set 2 were proposed. Based on this, an additional constraint on the square root of the dq-axis current error was introduced to analyze the distribution of zero voltage vectors under different error bands, further simplifying the control set and leading to the proposal of streamlined control set 3. [Results] The simulation and real-time experimental results showed that streamlined control set 1 achieved identical performance to the traditional control set, reducing the voltage vector sequence to 36 and decreasing the runtime to 53.31% of the traditional control set. Streamlined control set 2 and streamlined control set 3 showed performance comparable to traditional control set with their voltage vector sequences reduced to 9 and 4, respectively, and runtime reduced to 19.98% and 12.57% of the traditional control set’s runtime. [Conclusion] The three proposed streamlined control sets based on error constraints and voltage distribution patterns can significantly reduce traversal times while maintaining comparable control performance. This approach effectively optimizes real-time performance and provides a new solution for reducing the computational burden in two-step MPCC.