Vegetation is a key factor which comprehensively reflects condition of ecology and environment in mining area. With NDVI index as the parameter, vegetation coverage in Datong coal mining area from 1999 to 2010 were measured based on dimidiate pixel model by remote sensing estimation method. Spatial correlation indexes such as Global Moran's I, Getis-Ord General G, Getis-Ord Gi^* and Anselin Local Moran I were respectively used to reveal the temporal evolution and intrinsic mechanism of vegetation disturbed by coal mining from the global evolution and local effect perspective. Research results showed that: (1) Changing of vegetation coverage in Datong coal mining area presented cluster distribution and strong spatial autocorrelation in space, meanwhile, the phenomenon was absolutely obvious when vegetation coverage was in the lowest period. (2)In the rising period of vegetation coverage, hotspots of vegetation coverage gradually reduced or separated to the secondary hotspots area, clustering effect of vegetation changing became less significantly, which presented discrete state in general. (3) The disturbance of coal mining brought to vegetation showed overall characteristics and in proportion with coal production. (4)Different vegetation types presented different changing characters caused by coal mining.(5)The disturbance effect which large-scale coal mining of high efficiency and large production coal mines adopted circulation techniques brought to vegetation was still inconspicuous, however, the long-term effects need to be further studied.