In this article, a two-level methodology is proposed to optimize the design of a hybrid excitation synchronous machine (HESM) for a given electric vehicle (EV) over an arbitrary-selected driving cycle. We are looking at a huge analysis problem of finding an optimal hybridization ratio (HR) between the two excitation sources, namely, permanent magnet (PM) and wound excitation (WE). To find the optimal HR, the HR is scanned from 0 to 1 or from pure WE to pure PM excitation. For each HR, the motor is optimally designed at the component level, its cost is minimized, and its global efficiency over the selected driving cycle is calculated. Then, at the system level, the global efficiencies associated with each HR are compared in order to find the optimal HR. The complexity of the design optimization at the component level is addressed by nondominated sorting genetic algorithm II (NSGA-II). To make a compromise between the accuracy and speed of calculations, a nonlinear 3-D dynamic magnetic equivalent circuit (MEC) model is developed and evaluated by commercial finite element analysis (FEA) software. Following the proposed methodology and due to 300 h of computations with 48 CPU cores in parallel, the final HESM design can achieve up to 18.65% higher global efficiency than pure WE and 15.8% higher than pure PM excitation.