Abstract: A method for the structure based drug design, searching for and selection of potential medicinal compounds is proposed, which comprises predicting the value of the ligand binding affinities from the score calculated with the help of a scoring function with taking into account the protein structure, the ligand structure and the ligand position in the protein binding site. In the elaboration of the scoring function information about the already known both active and inactive ligands is employed. The use of the information about the inactive ligands makes the proposed method of elaborating the scoring function fundamentally different from all the known methods and allows not only to essentially improve the quality of the scoring function being elaborated, but also to constantly improve this quality as new experimental data become available.