The front-line antituberculosis drug isoniazid (INH) inhibits InhA, the NADH-dependent fatty acid biosynthesis enoyl ACP-reductase from Mycobacterium tuberculosis, via formation of covalent adducts with NAD (INH-NAD adducts). While ring tautomers were found the main species formed in solution, only the 4S chain INH-NAD tautomer was evidenced in the crystallized InhA:INH-NAD complex. In this study we attempted to explore the modes of interaction and energy binding of the different isomers placed in the active site of InhA with the help of various molecular modelling techniques. Ligand and enzyme models were generated with the help of the Vega ZZ program package. Resulting ligands were then docked into the InhA active site individually using computational automated docking package AUTODOCK 3.0.5. The more relevant docked conformations were then used to compute the interaction energy between the ligands and the InhA cavity. The AM1 Hamiltonian and the QM/MM ONIOM methodologies were used and the results compared. The various tautomers were found docked in almost the same place where INH-NAD was present as predicted by earlier X-ray crystallographic studies. However, some changes of ligand conformation and of the interactions ligand-protein were evidenced. The lower binding energy was observed for the 4S chain adduct that probably represents the effective active form of the INH-NAD adducts, as compared to the 4R epimer. The two 4S,7R and 4R,7S ring tautomers show intermediate and similar binding energies contrasting with their different experimental inhibitory potency on InhA. As a possible explanation based on calculated conformations, we formulated the hypothesis of an initial binding of the two ring tautomers to InhA followed by opening of only the ring hemiamidal 4S,7R tautomer (possibly catalyzed by Tyr158 phenolate basic group) to give the 4S chain INH-NAD tight-binding inhibitor. The predictions of ligand-protein interactions at the molecular level can be of primary importance in elucidating the mechanisms of action of isoniazid and InhA-related resistances, in identifying the effective mycobactericidal entities and, in further step, in the design of a new generation of antitubercular drugs.