Abstract: This paper introduces three main models(linear, logarithmic,and exponential model),which are used in monitoring the soil moisture at present according to the thermal inertial theory based on NOAA AVHRR data, and the processing steps of these models are presented in details.Three models are tested and validated by using ground truth data, then the environment which causes drought is also discussed in the paper by using ancillary data including DEM and landcover.The results show that three models meet the needs of the large scale drought monitoring, and that the application of ancillary data not only makes the quantitative analysis of drought distributing and its causes possible, but also validates the models correctness in a certain extent. The example,which uses NOAA AVHRR data and DEM and landcover data to monitor the drought condition in Hubei province on Nov. 12, 1998, gives that the terrain condition is also the main factor which affects the soil water content in southern China as well as the climate condition, and that the drought-against capacities of the different landcover types are of diversity. Forest is bot, woodland and grass take the second place and crop worst. Therefore, these conditions must be taken into account in monitoring drought in southern China.