In view of the frequent occurrence of extreme weather events in the world, the isometric quantile method was adopted to build the high-precision precipitation and temperature database of CMIP6. Coupled with Palmer drought index (PDSI), the spatiotemporal evolution of global and arid zone drought characteristics under future scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5) was estimated. The results show that the simulation accuracy of CMIP6 with bias correction is significantly improved. In the base period (1979-2014), the global drought frequency changes slightly, the global drought duration and drought intensity are relatively strong, and the drought intensity in the arid zone is relatively weak. During the forecast period (2025-2060), the global drought characteristics have significant differences in spatio-temporal distribution, with higher drought frequency, increased average drought duration and increased average drought intensity. In the forecast period, the frequency, duration and intensity of drought in the arid zone will increase, and the trend of drought in the tropical/subtropical arid zone and the temperate arid zone will become more obvious.