Dynamic wireless power transfer (DWPT) enables in-motion charging, reduces reliance on large onboard batteries, and mitigates range anxiety. During operation, longitudinal misalignment and vertical air-gap variation cause drift in both mutual and self-inductances, detuning the resonant network and degrading power regulation and efficiency. This work presents a multi-objective optimization of a detuned LCC-S compensation network to suppress output power fluctuation and total loss while maintaining zero-voltage switching (ZVS) across the coupling range. A unified analytical model characterizing the LCC-S behavior under inductance drift is developed, and a non-dominated sorting genetic algorithm (NSGA-II) is used to obtain compensation parameters balancing CP regulation and loss. A 270 W prototype validates the method: under 2× variation in mutual inductance and 6% variation in self-inductances, the output power remains within ±3.6% of the nominal value and the peak efficiency reaches 92.3%.
