Abstract: Under the combined effects of climate change, intensified human activities, and structural aging, dam safety deficiencies have long been a major concern in reservoir safety management in China. The dispersed and unstructured nature of dam defect information, limits its effective utilization. Existing understanding of dam safety hazards has mainly relied on expert judgment and limited engineering statistics, while nationwide quantitative analyses remain scarce. In this study, based on national dam inventory database and safety assessment reports of more than 20,000 deficient reservoir dams identified over the past decade, artificial intelligence techniques were used to extract and structure hazard information, resulting in the preliminary construction of a nationwide reservoir dam defect database. Statistical analyses were conducted from the perspectives of engineering scale, construction time, dam type, dam height, and geographic location to analyze dam safety hazard characteristics at the national scale, and corresponding countermeasures were proposed. The results show that seepage-related and structural safety issues are the two most prominent hazard types. Flood control deficiencies are most prominent in medium and small reservoirs, particularly for small reservoirs in northern regions such as the Songliao and Yellow River basins. Seismic safety hazards are relatively concentrated in the Haihe and Southwest River basins. Metal structure problems are more evident in large and medium reservoirs. The findings provide technical support for dam safety management, defective reservoirs rehabilitation decision-making and the establishment of a long-term governance mechanism.