The prediction of soil C stocks across the landscape has been increasingly studied in many areas of the world. Soil organic C storage in mountain areas is highly heterogeneous, mainly as a result of local-scale variability in the soil environment (topography, stoniness, parent material) and microclimate. The aims of the present study are to estimate soil organic C stocks (SOCS) in mineral soils of high-altitude grasslands of the Pyrenees and determine whether climatic and topographic variables can be used as predictors of SOCS and organic C content in the surface soil horizons of these ecosystems. For that purpose we sampled 35 soil profiles in subalpine and alpine grasslands including a range of altitudes, slopes and aspects. We analysed the soils for stoniness, bulk density, total C, texture, and C-to-N ratio and determined topographical variables. We used georeferenced climatic information for climatic descriptions of the sites. SOCS were highly correlated with soil depth. However, we were not able to predict soil depth by using environmental and topographic variables. In spite of this fact, altitude and aspect explained 41.2% of the SOCS variability while summer temperature and precipitation combined with aspect explained 56.9% of the variability of the organic C content of the surface layer (OC). The SOCS were low at high altitudes, probably as a result of an overall temperature limitation of net primary productivity. Under these conditions, the effect of aspect was small. The highest SOCS occurred at the lowest altitudes for ENE or WNW aspects, showing sharper decreases towards the south than to the north. The harsh climatic conditions and low-plant productivity that occur at the northern slopes reduced SOCS at the highest altitudes. In contrast, southern aspects showed similar organic C content along the altitudinal gradient. The OC variability in the surface soils not explained by climatic or topographic variables was partially related to the characteristics of soil organic matter, which may depend on the plant communities.