Abstract: In this paper the daily precipitation data from five global climate models (GCMs), which are all from CMIP5 and also the input in the Inter-Sectoral Impact Model Inter-comparison Project (ISIMIP), under RCP2.6 and RCP8.5 scenarios respectively in the Hanjiang River Basin is used to project the drought changes in the area through Standard Precipitation Index (SPI) at different time scales. Based on the observed data from 22 meteorological stations in Hanjiang River Basin in 1960-2004, and the GCM data in 2020-2059, three characters related to the drought events in these two durations, including drought severity, frequency and duration are compared and analyzed. Results show that in the drought severity the light or moderate drought events will happened less while the severe drought events can arise more often. As to the frequency and duration it is difficult to draw an identical conclusion from different GCMs. Anyway there is a general tendency in less frequency and shorter duration in the future. Through the comparison between the historical observed data and GCM generated data, HadGEM2-ES shows better performance in the rainfall simulation while GFDL-ESM2M and IPSL-CM5A-LR have better results in the less-rainfall namely drought analysis.