In this paper, a switching Model Predictive Control strategy is proposed for a Waste Heat Recovery system in heavy-duty automotive application. The objective is to maximize the WHR system output power while satisfying the output constraints under highly dynamic engine variations. For control design, a WHR system architecture with the expander and pumps decoupled from the engine is considered. Compared to a WHR system with the expander coupled to the engine, up to 29% more output power is obtained for the considered design. This holds for both steady state and highly dynamic engine conditions. The simulation results are obtained using a validated high-fidelity WHR system model with realistic disturbances from a Euro VI heavy-duty diesel engi...