Based on physical consideration of land surface energy balance, a theory is proposed for assessing relative soil moisture in the rooting depth with remote sensing data. The basis of the theory is the derivation of relative soil moisture in the rooting zone which is theoretically shown to be derivable from relative evaporation. The relationship derived between the relative soil moisture and relative evaporation is confirmed with experimental data collected with lysimeter measurements and in intensive field campaigns. Further it is shown that the proposed theory can be used to define a drought severity index (DSI) for drought monitoring, when the relative evaporation can be determined with remote sensing data. For this purpose, a demonstration in North China is performed. The used remote sensing data are NOAA/AVHRR which is available on a daily basis, the required meteorological data (wind speed, air temperature, humidity and pressure) are obtained from the operational measurement network of the National Meteorological Center of China. Comparisons between the estimated DSI and the actual measurements of soil moisture confirm the validity and robustness of the proposed theory.