Abstract: Using Landsat TM data, based on the spectral character analysis of main land-use types, different feature models under strong land disturbance in Antaibao opencast coal mine were set up. Vegetation higher cover area, vegetation lower cover area, denudation and cumuli area, excavation and transport area and slope area in Antaibao were extracted. The information of areas with higher vegetation cover and lower vegetation cover were detected by NDVI(normalized difference vegetation index). The information of higher vegetation-cover area was obtained by (TM4-TM5)>0, and compared with the vegetation cover area of NDVI. The information of excavation and transport area was obtained by TM4<40. The information of sloping area in range of 0.99～1.01 by TM4/TM7 was also obtained. Principal components analysis was applied for analyzing TM image. The characteristic information of denudation and cumuli area, excavation and transport area, and slope area is highlighted in the first principal components. The information of higher and lower vegetation-cover area is highlighted in the second principal component. Results show that the contribution rate of two principal components in the front achieves 97.16%. And object feature was classified by using characteristic of disturbance and vegetation. This technology and method can provide accurate data support for the dynamic monitoring of feature changing in opencast coal mine and the rebuilding and reclamation of land.