The occurrence and development of regional drought are affected by various water shortage processes such as surface water, soil water and crop water. The drought index based on a single water shortage process in the past is difficult to reflect the actual drought. It is of great significance to effectively carry out regional drought monitoring by comprehensively considering various water shortage processes to construct a comprehensive drought index. Based on the simulation data of hydrological crop coupling model (VIC-EPIC), a comprehensive drought index (CDSI) reflecting the combined water deficit of surface-soil-crop was constructed by linear weighted integrated standardized runoff index (SRI), soil water content anomaly index (SMAPI) and crop water deficit index (CWDI). The application research was carried out in Xiangjiang River Basin based on the data from 1981 to 2020. The results showed that the correlation between the drought characteristic sequence based on CDSI and the drought rate of 35 counties in the basin from 1990 to 2007 was between 0.4 and 0.8, and 82.9% of them passed the 5% significance level test. Compared with SRI, SMAPI and CWDI, which reflect a single water deficit, the correlation has been greatly improved. CDSI can better reflect the occurrence and development of regional actual drought, and can provide effective support for accurate identification and evaluation of large-scale drought.