Conventional model predictive control (MPC) requires receding horizon optimization and prediction of all voltage vectors in a threephase twolevel inverterdriven induction motor system, and has disadvantages of high computational cost and difficulty in implementation. To solve these problems, a deadbeat MPC is proposed, which reduces the receding horizon optimization prediction voltage vector from 8 to 3. The deadbeat voltage vector is calculated and the deadbeat sector is divided. The sector is determined based on the space voltage vector position. A singlestep receding horizon optimization prediction of the vectors in the sector is performed, and the corresponding optimal switch state is output. The simulation results show that the deadbeat MPC can achieve almost the same steadystate tracking performance as the traditional MPC while reducing the computational burden of the system.