This paper presents the design of non-linear torque control strategy for an engine, developed as the weighted sum of linear model predictive control (MPC) laws. The controller is optimized for tracking a reference torque signal and for reducing CO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> emissions. The engine model is taken from Toyota student competition of European Control Conference (ECC) 2015. Subspace identification techniques were employed to identify linear local models of the engine. Two control strategies are proposed: a) linear MPC and b) MPC weighted sum approach. Both solutions are compared to determine the best one. Lastly, MPC weighted sum control shows better results than linear MPC in tracking problems and the performance index proposed by the competition.