[Objectives] Bank collapse is a major form of planform deformation of alluvial riverbeds and one of the major natural disasters in the middle and lower reaches of the Yangtze River. However, due to multiple influencing factors and complex mechanisms of bank collapse, its accurate prediction and early warning remain challenging. After the construction and operation of the Three Gorges and upstream cascade reservoir groups, the Jingzhou section of the Yangtze River in Hubei Province shows long-distance and long-term scour trends, with significantly increased bank collapse risks, seriously affecting flood control, navigation, and socio-economic development along the river. This study aims to develop a method for predicting bank collapse under continuous scour conditions in the middle reaches of the Yangtze River, providing technical support for the establishment and practical application of a multi-indicator bank collapse risk assessment model. [Methods] A comprehensive bank collapse risk evaluation indicator system was developed for the middle reaches of the Yangtze River based on the analytic hierarchy process, encompassing three dimensions: current bank collapse status, substrate conditions, and near-bank variations, with a total of six characteristic indicators. On this basis, a comprehensive bank collapse risk assessment model in the middle reaches of the Yangtze River was established. Three-level early warning classification criteria for bank collapse were proposed, and they were applied to predict bank collapse risks in the Jingjiang and Honghu sections of the Yangtze River mainstream. [Results] Following the operation of the Three Gorges Project, most of the bank collapses and areas with high bank collapse intensity in the Jingzhou section of the Yangtze River mainstream were largely associated with local river regime adjustments. In addition to collapse occurring in unprotected bank sections, many failures occurred in the weak parts of protected sections or lightly protected sections, with a notable increase in sudden bank collapse events. In 2024, the Jingzhou section of the Yangtze River mainstream in Hubei had nine bank sections predicted to be at high risk of collapse with a red warning level. The majority of the high-risk bank collapse sections were distributed in natural unprotected sections, though some protected sections still had relatively high early warning levels for bank collapse. Among them, the total lengths of the Level I and II warning bank sections were 3.54 km and 16.76 km, respectively. [Conclusions] Based on the evaluation results of typical bank sections including protected, unprotected, and mainstream-adjacent banks, the bank collapse risk assessment model constructed in this study demonstrates certain applicability for bank collapse prediction in typical sections of the middle reaches of the Yangtze River. The characteristic indicators show certain sensitivity to variations in different bank conditions, and the proposed classification criteria for bank collapse early warning levels are reasonably sound.