Abstract: At present the methods using remote sensing technology to monitor soil moisture content become popular.This paper selected three monitoring methods which are widely applied with better stability and require less meteorological data,namely the SWEPDI index method,the energy index method and the TVDI index method,to better understand the characteristics of the landscape changes under the hinterland drought conditions in Mu Us Sandland,Inner Mongolia,China.Using Landsat8 data in April 2016 and September 2016 as data source,each of the three methods was conducted a linear fitting with the field survey data according to the time,the soil depth and corresponding soil moisture content.A comparison of the results from the models determined the more suitable model.At the same time,the change trend of soil moisture in 2 time phases was analyzed based on the plaque type by using the landscape index.The results showed that the effect of TVDI index method is superior to both the SWEPDI index method and the energy index method,and the TVDI index method also solved the problem that existed in continuously monitoring the soil moisture content.The TVDI index method is suitable for all kinds of vegetation conditions and various soil depth in drought monitoring.In addition,six landscape indexes were chosen to analyze the landscape pattern changes in April 2016 and September 2016 at different levels of drought,and it is found that the PLAND reached 58.76% in April,and it became 44.16% in September when it is at the light drought grade.Both were dominant as other indexes like LPI,AREA_CV and AI index all reached the maximum value.The drought condition was improved obviously.

Key words: SWEPDI index method, energy index method, drought index of temperature vegetation, landscape pattern