Abstract: Reliable control of the permanent magnet synchronous motor is necessary to ensure high speed-following capability and robustness under model parameter and load torque variations. This is often difficult to achieve using conventional linear controllers, as permanent magnet synchronous motor (PMSM) is a nonlinear and high coupling system containing many uncertainties. This paper proposes a nonlinear model predictive controller for a speed control of PMSM. The nonlinear PMSM decouples into a new linear system via the input-output feedback linearization scheme. To guarantee its convergence, an iterative quadratic program routine is proposed to solve the linear model based predictive control, problem with nonlinear constraints. Simulation results show the proposed controller has good dynamic and static performance and robustness under system parameter and load torque variations while reducing computational burden and torque ripples.