Mountain ecosystems will likely be affected by global warming during the 21st century, with substantial biodiversity loss predicted by species distribution models (SDMs). Depending on the geographic extent, elevation range and spatial resolution of data used in making these models, different rates of habitat loss have been predicted, with associated risk of species extinction. Few coordinated across-scale comparisons have been made using data of different resolution and geographic extent. Here, we assess whether climate-change induced habitat losses predicted at the European scale (10x10' grid cells) are also predicted from local scale data and modeling (25x25m grid cells) in two regions of the Swiss Alps. We show that local-scale models predict persistence of suitable habitats in up to 100% of species that were predicted by a European-scale model to lose all their suitable habitats in the area. Proportion of habitat loss depends on climate change scenario and study area. We find good agreement between the mismatch in predictions between scales and the fine-grain elevation range within 10x10' cells. The greatest prediction discrepancy for alpine species occurs in the area with the largest nival zone. Our results suggest elevation range as the main driver for the observed prediction discrepancies. Local scale projections may better reflect the possibility for species to track their climatic requirement toward higher elevations.