In this paper, a novel matching circuit design method utilizing a genetic algorithm and measured S-parameters of randomly moved coils is discussed. From the analytical comparison of different matching circuit topology, the superiority of active matching circuit is suggested, and potentially there is 21.4 % improvement in wireless power transfer efficiency by using 4 cell active matching circuit, which can create 16 different impedance values. Also, the matching circuit design simulation can be simplified by choosing representative impedance values of coils utilizing K-means clustering and use these values as targets. This drastically reduce the time for the matching circuit design simulation, especially for the matching circuit with larger number of cells.