Aiming at the problem of system performance degradation due to parameter error in permanent magnet synchronous motor control system, a robust current predictive control method based on parameter identification for permanent magnet synchronous motor was proposed. The sensitivity of the system to the inductance parameter was improved by introducing a robust current prediction method, but in the actual operation of motor, motor inductance parameter errors mainly influenced the daxis current response, and flux parameter errors mainly influenced the qaxis current response. At the same time, the stator flux parameters influenced the current steady state value in the constant acceleration process, even leading static difference between a given current and feedback value. In order to improve the performance of the current control system on the basis of robust predictive current control, recursive least squares method was used, the inductance, resistance and flux three parameters in the motor model were divided into two parts respectively for online identification, to reduce the complexity of the identification algorithm and dependence on processor performance. The simulation and experimental results show that the current predictive control method can further improve the accuracy of the current control system under the premise of ensuring the stability of the system.