This data release contains the associated data described in the related primary publication, “Predicting Flood Damage Probability Across the Conterminous United States” (Collins et al. [2022], see Cross Reference section). Publicly available geospatial datasets and random forest algorithms were used to analyze the spatial distribution and underlying drivers of flood damage probability caused by excessive rainfall and overflowing water bodies across the conterminous United States. Datasets contain input files for predictor and response variables used in the analysis and output files of flood damage probabilities generated from the analysis.