The study of ligand-protein interactions is of major interest in the discovery of new drugs. Accurate methods of calculating ligand binding (free) energies in a protein are useful for potentially speeding up the design of new ligands. During this thesis work, various strategies for the theoretical prediction of the affinity of the ligands towards the receptor will be implemented, very fast molecular mechanics (MM) employing classical force-fields, such as the molecular mechanics sampling coupled with Poisson-Boltzmann surface-area solvation (MM-PBSA) and / or free-energy perturbation (FPE) methods, up to the most complex, presumably more accurate and time-demanding hybrid quantum mechanics / molecular mechanics (QM/MM) approaches. QM/MM is a molecular simulation method that combines the strengths of the QM (accuracy) and MM (speed) approaches. For the identified leads the study of the binding processes simulations will be also carried out using the Enhanced sampling techniques. We presume that combination of the above “multi-level” protocols will efficiently guide the design of novel class of inhibitors of the XIAP protein (protein inhibitor of apoptosis linked to the X chromosome).