Studies on dietary functional responses in large herbivores are traditionally conducted by following individual animals. The method is very time-consuming, and hence, typically provides only a narrow array of forage species compositions. Here we use a range level approach to look at moose (Alces alces) selectivity for and utilization of forage species in relation to availability in both summer and winter. We compare 12 Norwegian ranges representing a large scale gradient in plant communities. The most important forage species in the diet were birches (Betula spp., comprising 43% of all trees browsed in summer and 27% in winter), rowan (Sorbus aucuparia, 25% of trees browsed in summer, 37% in winter), and bilberry (Vaccinum myrtillus, 42% of herbaceous epidermal fragments in summer feces). Selectivity for birches was positively related to its availability and negatively related to availability of rowan, Salix spp., and aspen (Populus tremula) together (all more selected for than birches). Multiple regression models including availability of several forage species were thus superior to single-species models in explaining the diet content of main forage plants. Selectivity for birches was also stronger in summer than in winter, while the opposite pattern was found for rowan. The finding is relevant for our evaluation of the quality of summer and winter ranges, and hence, their relative influence on population productivity. Our study underlines the need to incorporate species composition of available forage when quantifying dietary functional responses in selective herbivores such as moose. Furthermore, care should be taken when extrapolating data on moose diet across ranges or seasons.