Abstract: The surface subsidence caused by underground coal mining has formed a water accumulation area in Huainan
mining area,which significantly changes the distribution of surrounding soil moisture,and then affects crop growth and regional
food security. An improved temperature-vegetation drought index inversion method based on normalized kernel vegetation index
is proposed,which is applied to remote sensing monitoring of soil moisture content in coal mining subsidence area. The verification
results show that the accuracy of the inversion method is better than that of the traditional method,the sample correlation
coefficient is as high as 0. 75,and the monitoring error is significantly reduced. The spatio-temporal distribution map of soil
moisture level constructed based on this method reveals that the seasonal variation of water pattern in subsidence area is obvious.
The wet area is the main area in spring,the saturated area is dominant in summer,and the overall humidity tends to maintain
or improve in autumn and winter. On the whole,the proportion of medium and high moisture areas increased year by year,
which revealed the improvement of soil moisture recovery in the waterlogged area after subsidence. The research shows that the
inversion method has high applicability and stability in the dynamic monitoring of soil moisture in the mining area,and can provide
reliable technical support for ecological restoration evaluation and agricultural management.

Key words: coal mining subsidence,soil moisture content,vegetation index,drought index,temporal and spatial variation