Taking Jiangxi province as the study area，NDVI-Ts and EVI-Ts feature spaces are constructed using MODIS vegetation index and surface temperature remote sensing products，and the differences of feature spaces constructed based on different vegetation indices are compared and analyzed.Based on different vegetation indices，the TVDI and TVDI-E in the study area are calculated from August to October 2022 respectively，and soil moisture station data are used for correlation analysis of TVDI of different vegetation indices. The results show that the overall fitting effect of the feature space constructed based on EVI is better than that of NDVI；TVDI，TVDI-E and soil moisture site data all show significant negative correlations，in which the correlation between TVDI-E and soil moisture is higher than that of TVDI；TVDI-E is more applicable to drought monitoring in Jiangxi province. The results can provide a reference for drought monitoring research in Jiangxi province and similar areas.