Most current traction control systems use digital signal processing as their core, and their inherent latency and sequential instruction execution characteristics do not satisfy some applications that require higher performance. To solve this problem, it is desirable to use field programmable gate arrays (FPGA) with hardware parallelism and quasi-analog circuit characteristics to build high-performance control systems. In this paper, an FPGA implementation for permanent magnet synchronous motor (PMSM) control is proposed and a model-based design approach (Xilinx System Generator) is used to reduce the development time of the algorithm. First, the components and design ideas of each module of the vector control system are introduced. The PI, modulation, and coordinate transformation modules are introduced, and comparative simulation experiments are conducted to verify the dynamic and steady-state performance of the motor control system in agreement with Simulink simulation results. Next, the generated code is analyzed by post-implementation timing simulation, and the results show that the runtime of the FPGA-based program is significantly shorter than the runtime of programs in digital signal processors (DSP) and the generated code can achieve the correct functions. Finally, hardware resource consumption and design optimization are analyzed, and the results show that the optimized control system achieves improved performance and a balance with resource consumption. The above verifies the correctness and feasibility of the proposed implementation method.