Abstract: —Model predictive control (MPC) for multilevel converter has a multi-term cost function that requires high computation burden. In this paper, the deviation terms considered are classified into two categories according to their dependency on the voltage vector or the switching states of the multilevel converter. Subsequently, two approaches to reduce the computation burden are proposed. Mathematical models for the proposed MPC approaches are discussed and analyzed. An experimental platform for a fivelevel T-type multilevel induction drive is used to validate the proposed study. The experimental results show that the proposed approaches have substantially reduced the MPC computation burden compared to the conventional MPC techniques.