Abstract: In the field of high-performance permanent magnet synchronous motor servo control, the use of current prediction control strategy makes the current loop of the permanent magnet synchronous motor system have faster output response and smaller current ripple compared to traditional PI control. The digital system delay problem that exists during the current prediction control cycle is analyzed. A two-step current prediction control strategy is adopted to address the current prediction error caused by parameter mismatch in the current prediction model. A current prediction control model for permanent magnet synchronous motors is established, and a parameter identification strategy based on model reference adaptation is proposed. The current prediction controller design is completed. It avoids the tedious process of parameter tuning in traditional controllers, overcomes the digital delay of one beat of control current lag, optimizes the steady-state current control performance, achieves accurate and fast identification of motor inductance and flux parameters, and is of great significance for improving system robustness. The simulation results verify the feasibility and effectiveness of the proposed method.