In traditional deadbeat predictive speed control (DPSC) of permanent magnet synchronous motors (PMSMs), there is an issue with model parameter dependency. To enhance the control performance of PMSM speed regulation systems, this paper proposes a predictive speed control strategy based on adaptive parameter identification. First, the speed and current loops are designed according to the principles of deadbeat control. Then, mechanical parameters are identified using a model reference adaptive method, and these identified parameters are compensated in the deadbeat speed control. Finally, experiments are conducted to evaluate the proposed control strategy. The experimental results demonstrate that this approach improves the system's steady-state control accuracy and dynamic response speed.