Abstract: There is a lack of effective rainstorm disaster risk early warning models tailored to industrial agglomeration areas. In this study, a refined rainstorm disaster risk early warning model was developed based on meteorological observations, basic geographic and socioeconomic data for the Ningdong Energy and Chemical Industrial Base, a typical industrial park in Yinchuan City, Northwest China. The model was further validated using satellite remote sensing imagery. The terrain of the Ningdong Base is higher in the south and east and lower in the north and west, with population and economic activities mainly concentrated in its north-central area. The historical maximum cumulative precipitation reaches 203.2 mm, which may trigger rainstorm-induced hazards including factory waterlogging, road inundation and equipment failures. To better represent the unique characteristics of industrial parks, industrial-specific factors such as enterprise density, enterprise type and production-link vulnerability were incorporated to develop the rainstorm disaster risk early warning model with 14 evaluation indicators covering four dimensions: hazard, exposure, vulnerability, and disaster prevention and mitigation capacity. Case validation of the risk warming model based on GF-1 satellite remote sensing imagery shows that rainstorm hazard levels are positively correlated with the increase in actual water body area. High-comprehensive-risk zones are spatially consistent with regions with high exposure and vulnerability of disaster-bearing bodies. The model exhibits high accuracy and applicability. When applied to a rainstorm event in the 2025 flood season, it effectively captured potential rainstorm risks and supported meteorological authorities in issuing early warning information in advance, demonstrating its practical value and potential for broader implementation in industrial parks.