Balancing the number of grazing animals with the level of plant resources is a core issue in grazing management. Complete, full-coverage vegetation surveys are often used for this purpose, but these are expensive undertakings. We have presented a method to downscale information from regional sampling surveys by poststratification using a land cover map derived from satellite-based measures of reflectance values. This approach opens new prospects for landscape-level evaluation of productivity. We applied this method to eight grazing districts (19–245 km²) in Setesdal Vesthei, Norway, in 2006. Sheep densities in three of eight grazing districts of Setesdal Vesthei fluctuated above the estimated grazing capacity. We fitted 43 sheep with Global Positioning System collars in two contrasting grazing districts in 2007–2008 to assess their selection of the land cover productivity classes in the map used for poststratification. In the area with high vegetation coverage, sheep selection increased in areas with an overall higher productivity, supporting the main basis of the approach. However, in the grazing districts with lower vegetation coverage, selection was higher for areas of overall low vegetation productivity. The likely explanation is the presence of small areas of snow bed vegetation with high-quality forage in areas with a generally rocky surface. Our study provides a first step toward a grazing capacity evaluation to achieve a sustainable management of sheep on alpine ranges of Scandinavia, and our approach is likely applicable to other open alpine ranges in the northern hemisphere.