In this study, a new molecular docking method is presented to improve the docking accuracy. We introduce to docking design a concept of residue groups based on induced-fit and use K Score (a kind of PMF scoring function) to score the docking position. Genetic algorithm with the multi-population evolution and entropy-based searching technique with narrowing down space is used to solve the optimization model for molecular docking. To evaluate the method, we carried out a numerical experiment with 134 protein–ligand complexes of the publicly available GOLD test set. Through the comparison with other popular docking software, the proposed method showed the higher accuracy. The average computing time of this study is 44.1s, that made it has advantages in the virtual screening. Among more than 61% of the complexes, the docked results were within 2.0 Å according to Root-Mean- Square Deviation (RMSD) of the X-ray structure.