At present, the research on coal pollution emissions has focused on plain areas and has largely ignored mountainous areas.In this paper the mountainous region of Mentougou in Beijing was selected as the research study area as it has 98.5% of its area covered by mountains. A stereo model was constructed to extract a DEM (Digital Elevation Model) using forward-looking and backward-looking ZY-3 images. Classification and a regression tree model (CART) combined with an object-oriented method were used to extract residential areas from GF-2 remote sensing images. Imagery interpretation along with statistical records of the residential areas were used to derive a coefficient of burning coal in a mountainous area. A statistical model was then developed to estimate coal consumption in a mountain area. A literature search helped us to determine emission factors and to estimate the specific emissions of PM₁₀, PM_2.5, NO_x, SO₂, and CO. With the help of remote sensing and a DEM this study demonstrates that the emission of coal-fired pollutants in mountainous areas can be measured efficiently and effectively. Terrain had a significant effect on coal combustion in winter. The derived coefficient of burning in a deep mountainous area was 12.5kg/m² while the shallow mountain area exhibited a value of 9.375kg/m². In the mountain region of Mentougou in 2017 the area of coal-fired residential buildings was 5.68km². The derived coefficient of burning in a deep mountainous area was 12.5kg/m² while the Shallow mountain area exhibited a value of 9.375kg/m². The coal consumption was 0.0652 million tons in the winter, but the emission of air pollutants varied greatly among the towns in the mountainous areas.