A Geographic Information System with a participatory approach was applied to evaluate the grazing resources of a mountain forest in central Mexico. The study region was enclosed within a protected area. Land was communally owned and sheep grazed. The region was classified according to the vegetation type using satellite imagery. In each class, biomass production, botanical composition and growth patterns of the understorey vegetation were measured during the rainy season. A participatory study was carried out with breeders to identify forage species, estimate flock sizes and locate the grazing distribution patterns on maps and aerial photographs. These data were linked with the biophysical properties of the landscape utilising a GIS. Finally, an analysis at landscape level was carried out to assess current stocking rates and identify areas prone to overgrazing. This helped to optimise grazing patterns and to delineate a buffer belt for the grazing-exclusion zone of the protected area.