To reveal the disaster-causing characteristics of extreme precipitation and its spatial risk distribution, a case study of the Fuhuan River Basin in the middle reaches of the Yangtze River was conducted. Based on daily gridded precipitation data from 1961 to 2022, nine extreme precipitation indices (PRCPTOT, Rx1day, Rx5day, SDII, R10, R20, R95p, R99p, and CWD) were calculated. Then, their spatiotemporal evolution characteristics were analyzed by Mann-Kendall abrupt change and trend tests, the moving T-test, and Theil-Sen estimator. Principal component analysis (PCA) was used to extract comprehensive indicators of extreme precipitation, run theory was employed to identify the disaster-causing duration and severity of extreme precipitation events, and non-dominated sorting was applied to classify risk levels in the basin. The results indicate that except Rx5day, all indices showed an increasing trend over time, with intensity indices exhibiting a spatial pattern of “increase in the downstream but decrease in the upstream”. The cumulative contribution rate of the first two principal components exceeded 77%, demonstrating an excellent performance on characterizing multidimensional features of extreme precipitation with integrated information of severity, frequency, and persistence. A “high-severity-long-duration” disaster pattern was pronounced in the downstream region, resulting in prominent risk. The disaster-causing risk was primarily influenced by frequent and accumulative moderate-severity precipitation processes characterized by indices such as Rx5day, R95p, and R20. The assessment framework developed in this study comprehensively reflects the prolonged and cumulative effects of disasters caused by extreme precipitation, providing a scientific support for watershed-scale flood risk management.