Most habitat models developed for defining priority conservation sites target areas currently exhibiting suitable habitat conditions. For species whose habitats have been altered by land use practices, these models may fail to identify sites with the potential of producing suitable habitats, if management practices were modified. Using capercaillie Tetrao urogallus as an example, we propose a model for evaluating the potential of ecological conditions at the landscape level to provide suitable habitat at the local scale. Initially, we evaluated the influence of selected landscape parameters on the structural characteristics of vegetation relevant to capercaillie. Then we used capercaillie presence data and an ecological niche factor analysis (ENFA) to identify landscape and land use variables relevant to capercaillie habitat selection. We also studied the effect of scale on predictive model quality. Despite high variance, correlations between landscape variables and forest structure were detected. The greatest influence on forest structure was recorded for climate and soil conditions, which were also found to be the best predictors of capercaillie habitat selection in the ENFA. The final model, retaining only two landscape variables (soil conditions and days with snow) and three land use variables (proportion of forest, distance to roads and forest-agricultural borders), explained a high degree of capercaillie habitat selection, even before considering patch size and connectivity. By restricting the analyses to areas with stable subpopulations and a set of relatively stable landscape variables capable of explaining habitat quality at a local scale, we were able to identify areas with long-term relevance to conservation of capercaillie.