A model of the interaction of substrates and inhibitors with xanthine oxidase (XO) based on similarity concepts and molecular modeling is introduced and discussed, and previous literature is reexamined in the light of recent insights into the mechanism and structure of XO. Use is made of quantum-chemical calculations with the inclusion of solvent effects, molecular superimposition with least-squares fitting algorithms, and molecular electrostatic potentials. First, the relative stabilities of the tautomeric forms of the physiological substrates, xanthine and hypoxanthine, are calculated both in vacuo and in water in order to select the most abundant form(s) at physiological pH: the two substrates prove to be stable in their lactam forms, with a dominance of the N7-H tautomer for xanthine and of N9-H for hypoxanthine. The structures of xanthine and hypoxanthine are then superimposed, and their relative orientation with respect to the molybdenum center of XO is suggested. The criteria used for superimposition reflect the importance of functional groups of xanthine and hypoxanthine, as inferred from experimental work. In particular, the carbonyl oxygen common to the two substrates is given special consideration on account of its determinant role. The results show that the most important functional groups of the two substrates can be successfully superimposed by means of a rotation that exchanges the five-membered with the six-membered rings of xanthine and hypoxanthine with respect to molybdenum. The close similarity of the electrostatic potentials of the two superimposed molecules adds weight to the proposed orientation of the substrates in the binding site. The model of interaction is then tested and further developed using a series of previously-synthesized dimensional analogs of xanthine and hypoxanthine. The results confirm that the correct positioning of the carbonyl group is essential if a productive interaction with XO is to be achieved and allow us to map the dimensions of the active site starting from the superimposition of the physiological substrates. Two hypotheses regarding the amino acid residues interacting with the important carbonyl oxygen of the substrates are then put forward on the basis of spectroscopic and biochemical evidence: they are postulated to be one lysine or one protonated glutamic acid residue. In an attempt to unify the binding of substrates and inhibitors, the model is extended to the inhibitors of XO by superimposing the most interesting inhibitors developed by Robins on xanthine and hypoxanthine. This allows us to define the most suitable location of the phenyl rings of these inhibitors with respect to the superimposition of the substrates. Intriguingly, the superimpositions of the most active inhibitors are consistent with a unique location of their phenyl rings, even though they are in different positions on the purine ring. Finally, the flavone, which is a potent inhibitor of XO and is currently under investigation by the authors, is accounted for by these findings and successfully included in the model. This model incorporates many important insights into XO and can be of general interest. Moreover, it represents a clear-cut alternative to a previous model developed by Robins on the basis of the coordination of substrates and inhibitors to molybdenum.