The Future_Global_AI_PET Database offers high-resolution global estimates (30 arc-seconds) of potential evapotranspiration (PET) and aridity index (AI), derived from 22 CMIP6 Earth System Models. It covers two historical periods (1960–1990; 1970–2000) and two future periods (2021–2041; 2041–2060), analyzing climate trends across four shared socio-economic pathways (SSPs). The dataset includes three multimodel ensemble averages (All, Majority Consensus, High Risk) to account for different levels of climate model uncertainty. A methodological overview, geospatial implementation details, and technical evaluation accompany the database. Historical data validation compared results with weather station data (PET: r² = 0.72; AI: r² = 0.91) and the CRU_TS v4.04 dataset (PET: r² = 0.67; AI: r² = 0.80). Given the anticipated climate shifts, this database serves as a valuable resource for scientific and practical applications, offering insights into predicted eco-hydrological and terrestrial ecosystem impacts.