Abstract: Recently, docking has been widely used to predict the binding-modes of protein-inhibitors, when the crystal complexes structure was absent. Most docking algorithms are able to generate a large number of probable conformations, it, however, is difficult to effectively evaluate these docking poses and identify the most reasonable binding-mode. In the present study, on the basis of the crystallographic data of human 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGR), a large number of docking statins poses were generated by using Surflex-dock and Autodock program. Then, the binding energy of human HMGR complexed with the corresponding docking statins poses were systematically investigated by using ONIOM and eXtended ONIOM methods. Our present calculation results indicate that the present scoring strategy can effectively find the favorable binding conformation (similar even close to the crystallographic complexes structure) from a large number of docking poses.