Aiming at the problem of slow convergence during the parameters identification of Permanent Magnet Synchronous Motor (PMSM), a combined algorithm based on orthogonal projection algorithm and multi-innovation recursive least square algorithm is proposed to estimate the parameters of PMSM. The mathematical equation of the four-order nonlinear state space model of PMSM is selected, and the model is rewritten as a linear regression model. The noise of motor operating system is added into the simulation process to simulate a real motor operating environment, and then the combined algorithm is applied in the identification of PMSM in the form of linear regression. The results show that the combined algorithm is effective. This work is supported by National Natural Science Foundation of China (No. 61572237).