The expansion of the wolf population all over Europe has posed problems on wolf-man coexistence in those areas where the wolf was not present for a long time. Breeding activities are now threatened by high predation on livestock. We investigated the relationship between wolf and its habitat, in order to evaluate wolf habitat suitability and to predict its presence. The study areas covered 3289 km2. A 23 km2 grid was used to identify 143 sample squares. For each sample square 58 habitat variables were measured from land-use (1:25 000), and from topographical maps. Wolf presence (4 classes) was assessed by scat collection, direct observations, snow tracking, wolf-howling, and predation records. Univariate and multivariate statistical analyses were carried out to investigate the influence of habitat features on wolf presence. The habitat suitability model was based on two different equations: a dichotomous logistic regression model, and a polytomous one. The first one discriminated between suitable and unsuitable habitats, and its predictions were confirmed in 93% of grouped cases; the second provided predicted values of wolf presence that were concordant with the observed response levels in 82.5% of contrasts. Both models underlined the importance of three factors in determining wolf habitat suitability: wild prey abundance, human presence and forest cover.