Coarse mode aerosol optical depth (cAOD) is critical for understanding the impact of coarse mode particles on climate, such as dust. Currently, the limited data length and high uncertainty of satellite products impair the applicability of cAOD for climate research. Here, we propose a spatiotemporal co-action deep learning model (SCAM) for the retrieval of global land cAOD (500 nm) from 2001 to 2021 at daily temporal resolution and 0.5° spatial resolution. The cAOD products are uploaded in Geotiff format, stretched from -89.5° to 89.5° latitude and from -179.5° to 179.5° longitude. On certain days the products might be unavailable, due to the missing MODIS satellite data used for calculating cAOD.