Soil moisture is the key element in study of underground coal mining disturbance in desert mining area. The Shendong Mining area, a typical desert mining area, and was selected as the research region. Temperature vegetation dryness index (TVDI), acquired from the bi-parabolic NDVI-Ts
space, was applied to monitor the soil moisture conditions in Shendong mining area based on the MODIS data from 2000 to 2015. The Gradient-based structural similarity (GSSIM) method was applied to quantitatively analyze the spatial distribution of soil moisture in Shendong mining area over the past 16 years. The results show that the change of soil moisture has distinct temporal and spatial heterogeneity characteristics in the Shendong mining area, specifically expressed as follows:(1) there was a gradually increasing trend of soil moisture from the northwest to the southeast; while the drought area dropped from 96.03% in 2000 to 59.59% in 2015;(2) soil moisture mutated in most of mining area, accounting for 60.08% of the total area from 2000 to 2015, and 49.87% of the vegetation cover in the mutated area showed significant betterment, as the soil moisture apparently improved. About 35.18% of the area showed changes in soil moisture while 28.13% of the vegetation cover improved with increased soil moisture. Soil moisture, in only 4.75% of the mining area, did not change significantly. The spatio-temporal distribution of soil moisture are influenced by landforms and underlying surfaces.