In this article, the three-phase asynchronous motor is taken as the research object, and the deadbeat predictive control is the control core, and its control performance is studied. Considering the overshoot of flux linkage and speed response in deadbeat predictive control, an improved deadbeat predictive control algorithm based on synergy theory is proposed. Different from deadbeat predictive control, improved deadbeat predictive control can control the flux linkage and speed independently through two adjustment parameters, which are achieved through the manifold of synergy theory. To accurately obtain the feedback flux linkage and load torque in the control, a flux observer and a load torque observer are designed. Based on these, a simulation platform is built, and the control performance of the two control algorithms is compared. The robustness of the two control algorithms is further verified by changing parameters. Finally, the experimental platform of three-phase asynchronous motor is built in this article. The experimental program based on deadbeat predictive control is designed. The no-load experiment and loading experiment are carried out, respectively. The feasibility of the algorithm is verified by experimental data.