As global warming intensifies, the risk of rainstorms and flooding in the Loess Plateau (LP) increases considerably. Accurately assessing the risk of rainstorms and flooding helps improve public's knowledge and ability to cope with disasters. In this study, we adopt multi-source data fusion and quantitative assessment techniques to construct a comprehensive assessment framework based on disaster system theory, achieving a quantitative assessment of disaster risk and representation of spatial distribution. Finally, the flood disaster risk for 2020-2100 was assessed using the comprehensive assessment framework. The results suggested that (1) The general trends in the spatial distributions of annual precipitation, flood season precipitation, and rainstorms over the historical period were similar. Overall, there may be an increasing trend of precipitation in the future. The intra-annual distribution of precipitation tends to even out as precipitation increases. (2) The flood disaster risk on the LP has a solid regional nature, with a decreasing pattern from the southeast to the northwest. The high and second-high flood disaster risks were mainly distributed in the central, southern, and southeastern. (3) The flood disaster risk in the LP from 2020 to 2100 induced by climate change and socioeconomic development would increase under combination scenarios of SSP1, SSP2 and SSP5, especially in the northeastern part of the LP. The study assessed the risk of flood disaster from past and future perspectives, which can provide the valuable reference for flood disaster risk management.