This dataset contains 6 categories and 38 attributes, providing detailed descriptions of DFBD events. This dataset records basic information (location, formation date, etc.), dam characteristics (height, length, volume, etc.), lake characteristics (area, capacity, length, etc.), debris flow characteristics (velocity, flow, volume, etc.), dam failure characteristics (peak flow, casualties, etc.), and climate characteristics (precipitation and temperature). The dataset shows that DFBDs have key characteristics such as instability, complete blockage, and overflow failure. Especially in China, the number of such dams has significantly increased, with 15% of rivers experiencing repeated mudslides, resulting in 35% of all DFBDs. Further analysis shows that in the stability assessment of DFBD, the Ls (AHV) model should be given priority, followed by the DBI model. Compared with other obstacle dam datasets, this dataset is more targeted, emphasizing the review of raw data and the unity of terminology and concepts (i.e. blockage patterns and stability) to ensure data consistency and accuracy. The dataset and results presented in this study may contribute to a deeper understanding of the formation, distribution, and evolution of DFBD.