Abstract The development of advanced control systems is motivated to achieve the complicated aircraft engine control, so a novel semi-alternative optimization strategy based model predictive control is proposed. In this proposed controller, a nonlinear state-space model is generated as the predictive model on the basis of a baseline engine model that estimates the current operating state of engine with an extended Kalman filter every sampling instant, and a quadratic constrained problem rather than a higher order nonlinear problem is constructed based on this model. Moreover, a novel control sequence to be optimized based on semi-alternative optimization strategy is proposed, which includes two parts. The first part represents that all the control variables of the engine are optimized for the current sampling instant while the second part represents that different control variables are optimized alternatively for different future sampling instants over the control horizon. The proposed model predictive control method is implemented to a twin-spool turbofan engine. Simulation results demonstrate not only that the proposed method can achieve a smaller control error than the standard model predictive control algorithm does, but also that the proposed controller is more time-saving than the standard model predictive control, which can save up to about 61% optimization time when control horizon increases to nine.