The evolution of drought events is a dynamic process with spatio-temporal continuity. Analyzing the migration characteristics of drought events from a three-dimensional perspective is significant to objectively identify the nature of droughts and improve our monitoring and projection capabilities. Existing three-dimensional drought identification methods focus on utilizing a simple overlapping logic to characterize the spatial and spatio-temporal continuity of drought patches, which ignores the dynamic of drought evolution. Here, we propose a three-dimensional drought identification method that considers both spatial autocorrelation and anisotropy. Subsequently, the spatio-temporal characteristics and migration patterns of meteorological drought events in China were identified during a historical period (1961-2010) and a future period (2031-2080, SSP2-4.5 and SSP5-8.5) using the ERA5 dataset and the multi-model ensemble mean (MEM) of CMIP6. The results suggest that, in the future, the number of meteorological drought events will decrease by over 70% in comparison with the historical period, but the severity, duration, affected area, and distance will increase significantly. The majority of future meteorological drought events will occur during spring and summer, specifically 96.3% under the SSP2-4.5 scenario and 95.0% under the SSP5-8.5 scenario. Future meteorological drought events will migrate predominantly to the north-east, with 33% (SSP2-4.5) and 38% (SSP5-8.5) being shown to do so. The hotspot area of the migration trajectory of meteorological drought events will shift from the upper Yangtze River Basin to the upper Yellow River Basin. These findings will help us implement drought prevention strategies and allocate water resources.