Abstract: The Hulunbuir meadow steppe, located in northeast China, is an important site for animal husbandry. The potential to further develop animal husbandry in this region is closely related to the growth of grass and to sustainable utilization of grasslands more generally. As a critical component of the grassland ecosystem, the intensity of grazing has become an important issue. In this study, remote sensing data, based on Landsat images, and Net Primary Productivity (NPP) data were collected from June to July 2014 and used to estimate grazing intensity in Xeltala pastures. Multi-image Landsat optical data was used to calculate aboveground biomass and biomass increments, while the NPP data was used to analyze grass growth. Unlike previous studies, this analysis used NPP data to take account of the heterogeneity of grassland conditions. The results indicate that this method is capable of accurately estimating grazing intensity, with a R² of 0.7996 when validated by measurements on the ground. Grazing intensity in the Xeltala pastures ranged from 1 to 2.5 Au/ha, with the latter being over-grazed. The local map suggested that heavy grazing areas were distributed around a small lake and in certain locations where abundant grass favoured concentrations of cattle. Light grazing areas, on the other hand, were located in enclosed places where the grass was reserved for winter supplies. Extending the study area to include the city of Hailaer, the grazing intensity map clearly demonstrated the variations between areas of light, moderate and over-grazing, with Hailaer in the southwest appearing as extremely lightly grazed compared to the northeastern meadows.