This study presents an implementation of the protein-ligand docking program GOLD and a generalizable method to predict the binding site and orientation of potential vanadium drugs. Particularly, theoretical methods were applied to the study of the interaction of two (VO)-O-IV complexes with antidiabetic activity, [(VO)-O-IV(pic)(2)(H2O)] and [(VO)-O-IV(ma)(2)(H2O)], where pic is picolinate and ma is maltolate, with lysozyme (Lyz) for which electron paramagnetic resonance spectroscopy suggests the binding of the moieties VO(pic)(2) and VO(ma)(2) through a carboxylate group of an amino acid residue (Asp or Glu). The work is divided in three parts: (1) the generation of a new series of parameters in GOLD program for vanadium compounds and the validation of the method on five X-ray structures of (VO)-O-IV and V-V species bound to proteins; (2) the prediction of the binding site and enantiomeric preference of [VO(pic)(2)(H2O)] to lysozyme, for which the X-ray diffraction analysis displays the interaction of a unique isomer (i.e., OC-6-23-Delta) through Asp52 residue, and the subsequent refinement of the results with quantum mechanics/molecular mechanics methods; (3) the application of the same approach to the interaction of [VO(ma)(2)(H2O)] with lysozyme. The results show that convenient implementation of protein-ligand docking programs allows for satisfactorily reproducing X-ray structures of metal complexes that interact with only one coordination site with proteins and predicting with blind procedures relevant low-energy binding modes. The results also demonstrate that the combination of docking methods with spectroscopic data could represent a new tool to predict (metal complex)-protein interactions and have a general applicability in this field, including for paramagnetic species.