As one of the major terrestrial ecosystems, grasslands play a vital role in the global carbon cycle. However, the estimation of carbon stock in China's grasslands still remains controversial. Using the measured data from spatial stratified sampling (including 200 sites; 400 soil profiles, 2400 soil samples in the depth of 0-100 cm, 400 above-ground biomass carbon data, and 2400 below-ground biomass carbon data), together with the EVI (enhanced vegetation index), the estimation of total carbon stock was calculated at 29.9 Pg, with an area of 2.63 x 10(6) km(2) and the storage of SOC was larger than that of biomass carbon regardless of grassland types. The carbon density of grasslands showed a tendency of decreasing from the southeast to the northwest. Alpine grassland contributed the most (74.2%) of the total carbon storage, while the Desert grassland contributed the least (4.3%). The vertical distribution of carbon density varied strongly among different grassland types and a proportion of 15.8% of total carbon still stored at the layer of 60-100 cm. Since this study combined spatial stratified sampling (based on grassland type and community level, sampled belowground samples to a depth of 0-100 cm) with remote sensing data, it is believed that this estimation of the stock and spatial pattern:of carbon in grasslands of northern China is more accurate. Therefore, these results will be conducive for understanding the contribution of different grassland types to the global carbon cycle and providing a reference for future measures taken for increasing carbon storage in grasslands.