The state-of-the-art MODIS surface reflectance products suffer from temporal and spatial gaps, which make it difficult to characterize the continuous variation of the terrestrial surface. We proposed a framework for generating the first global 500 m daily seamless data cubes (SDC500), covering the period from 2000 to 2022. We believe that the SDC500 dataset can interest other researchers who study land cover mapping, quantitative remote sensing, and ecological science.