Abstract: In view of the urgent need for meteorological disaster prevention and mitigation as well as the shortcomings of existing technical links, this paper presents an improved rainstorm flood simulation algorithm. The impact of previous precipitation, spatial and temporal distribution of precipitation, and the land surface characteristics were introduced in the algorithm. Systematic analysis of precipitation spatial interpolation, parameter optimization of rainfall runoff model, local parameterization of twodimensional hydrodynamic models, model coupling and applications were carried out. The algorithm can reflect and simulate the linkage of precipitation -runoff -inundation, which has been verified and applied in the Huaihe River basin. Results indicate that the spatial patterns of precipitation obtained by different spatial interpolation methods are generally consistent, and the Kriging method is preferred according to the cross-validation results. After optimizing the CN coefficient, initial loss rate and other parameters, the SCS model is able to simulate the relationship between precipitation and runoff. The surface roughness coefficients of the FloodArea model were determined based on land use type. Better simulation can be achieved by optimizing the FloodArea model for precipitation input. The coupling of the SCS model and the FloodArea model are realized by taking the rainfall runoffprocess as the link. The algorithm has significantly improved the simulation capability of rainstorm flood.