Surface incident solar radiation (Rs) is a key component of the surface radiation budget. It drives the global climate system and impacts the global energy balance and the hydrological and carbon cycles. Great progress has been made in the detection of variations in surface solar radiation (Rs) from meteorological observations, satellite retrieval and reanalysis. However, each type of estimation has its advantages and disadvantages. It has been shown that sunshine duration (SunDu)-derived Rs data can provide reliable long-term Rs variation over China; however, these data are spatially discontinuous. Therefore, we merged SunDu-derived Rs data with satellite-derived cloud fraction (MODAL2 M CLD) and CERES SYN aerosol optical depth (AOD) data to generate Rs data by the geographically weighted regression method. This dataset provides the monthly Rs from 2000 to 2017 over China with the spatial resolution of 0.1°.