In the present study, a hypothesis is presented for the prediction of the regioselectivity of cytochrome P-450 catalyzed hydroxylation of fluorobenzenes. The regioselectivity of the in vivo hydroxylation of fluorobenzene, 1,2-difluorobenzene, 1,3-difluorobenzene, 1,2,3-triluorobenzene, and 1,2,4-triflurobenzene could be predicted within 6% accuracy on the basis of the substrate's frontier orbital characteristics for electrophilic attack. The in vivo regioselectivity of the hydroxylation of fluorobenzene was not significantly influenced by changes in the cytochrome P-450 enzyme pattern. This implies that the regioselectivity is not predominantly determined by the juxtaposition of the relatively small substrates in the active sites of the cytochrome P-450s catalyzing the reaction. Additional in vitro experiments using 1,2-difluorobenzene as the model substrate demonstrated that minor factors influencing the regioselectivity and possibly responsible for the 6% deviation from the calculated values in in vivo experiments might be (i) the influence of biotransformation routes occurring in vivo but not of importance in in vitro microsomal incubations and (ii) a small variation due to influences of the contribution of various cytochrome P-450 enzymes. On the basis of the results obtained, it is concluded that the aromatic hydroxylation of fluorobenzenes proceeds through an initial electrophilic attack of (FeO)3+ on the aromatic substrate, and not through initial electron abstraction followed by attack of the (FeO)2+ species on the substrate radical cation. The fact that the regioselectivity observed could be predicted and/or explained by the site of initial (FeO)3+ attack also argues against epoxides as important intermediates in the formation of phenol metabolites from fluorobenzenes.