Background: Protein-ligand interaction plays a key role in controlling biological processes.Thus the computation of protein-ligand binding energy is a crucial step in computer aided drug discovery.Due to the huge amount of computational tasks in virtual screening and the complexity of protein-ligand systems, scoring functions with a lot of approximations are used other than accurate free energy approaches.This lets to low hit-to-lead rate.Methods: Quantum chemistry computations with new Gaussian basis sets and molecular dynamics simulations with the force field of drug molecules were used to improve the accuracy of protein-ligand binding energy computation.Scaling approaches were also used with respect to the ligand binding selectivity and molecular properties of drug-like compounds.Results: The accuracy ofprotein-ligand binding energy computation can be improved by introducing quantum chemistry correction and proper scaling.Conclusions: Higher accuracy of protein-ligand binding energy computation will be used to improve the efficiency of computer-aided drug design .