Land surface temperature (LST) was obtained by MODIS data and observation data from automatic weather station on the ground. Modify temperature-vegetation dryness index (MTVDI) was studied by LST combined with MODIS-EVI. Modified energy index (MEI) was studied by channel 2 data and LST test. Vegetation coverage was introduced, and modify perpendicular drought index (MPDI) was studied by channel 1 and channel 2 data. Combined with the soil moisture data from the agro-meteorological regular observation business, statistical functions of the indexes and soil moisture were established, which were all pass the 0.01 significance test. The models had been applied to Ningxia drought monitoring business, the results showed that the monitoring effects of MTVDI and MPDI were significant in crop growing season, and MEI, PDI were effective on bare ground or sparse vegetation. Each model has its own advantage and disadvantage, and the comprehensive application of the models could achieve the best results during the actual monitoring. The accuracy could reach to 90%.