In this paper, the magnetic coupler composed of two planar spiral coils was studied and optimized. In view of the LCC-S type WPT system, a multi-objective optimization method for geometric parameters of the magnetic coupler was proposed. Taking the coil radius, transmission distance and coil turns as decision variables, with the transmission efficiency, transmission power and total harmonic distortion as objective functions, the performance of the MC-WPT system was optimized by using a fast and elitist multi-objective genetic algorithm, that is, non-dominated sorting genetic algorithms II (NSGA-II). The Pareto solution sets of three objective functions were obtained. Finally, the optimized parameters were brought into the MATLAB/Simulink simulation model, and an experimental prototype was built. The simulation and experimental results verified the feasibility and effectiveness of the proposed method.