In recent years, serious drought-related disasters in China have received increased attention due to their significant impacts. Assessing drought variations was critical to ensure agricultural activity, social development, and human health. This study used the three-dimensional clustering method to investigate variations of drought events in historical (1971–2010) and future (SSP126 and SSP585, 2061-2100) periods in China based on observations and model simulations from Coupled Model Inter-comparison Project Phase 6 (CMIP6). The variations and risk analysis of drought characteristics were assessed by copula-based methods. Results indicated that precipitation was projected to increase by 12.0% and 19.5% for the SSP126 and SSP585 scenarios, respectively, with substantial increases in arid and semi-arid regions. The drought duration, severity, and affected area are shown to increase based on the CMIP6 projections compared with historical periods. We also assessed drought return periods of multiple drought characteristics for each scenario by using the copula-based joint distribution. Based on the joint analysis of duration, severity, and affected area for meteorological drought, both the “or” and “and” return periods in the future indicated increased risks. Under global warming, extreme drought events with long duration, high intensity, and extended affected area were shown to occur more frequently in the future, especially under the SSP585 scenario. These results can aid the understanding and characterization of drought events for drought risk management under global warming.