The Global L-band equivalent AI-Based Vegetation Optical Depth (GLAB-VOD) dataset provides a long-term, temporally consistent record of vegetation optical depth from 2002 to 2020 at 25 km resolution and 18-day intervals. Using machine learning, this dataset extends the SMAP-IB VOD record by fusing brightness temperatures from SMOS, AMSR-E, and AMSR-2 with ancillary soil moisture data from CASM. A companion product, GLAB-TB, offers harmonized brightness temperatures used as intermediate inputs. GLAB-VOD is designed for applications in vegetation monitoring, biomass estimation, and climate studies.