Abstract: Against the backdrop of intensified global climate change and frequent extreme precipitation events, traditional flood prevention models are facing severe challenges. As one of the countries most severely affected by flood disasters, China has actively promoted the deep integration of new technologies such as artificial intelligence (AI), the Internet of Things, and digital twins with flood prevention systems in recent years. Flood prevention work has shifted from being "experience driven" to "datadriven". The system has reviewed the evolution of AI technology in the field of water conservancy, from early expert systems to current deep learning and digital twin technologies, and analyzed its core breakthroughs in flood forecasting, urban waterlogging warning, and regional water network collaborative management. Research has shown that AI hybrid models can effectively improve the accuracy of peak flood forecasting; The urban waterlogging monitoring system supported by computer vision and Internet of Things technology can achieve "second level recognition minute response" and improve disposal efficiency by more than 6 times; The digital twin technology constructs a closed loop of "forecast warning rehearsal contingency plan" through virtual real interaction deduction, optimizing the efficiency of cross basin scheduling. However, the current intelligent flood control still faces challenges such as data quality, extreme event generalization, and cross departmental collaboration. In the future, it is necessary to deepen the integration of physical models and AI, build cross domain collaboration platforms, and expand the value of flood control data for people's livelihood services, in order to comprehensively enhance the resilience and intelligence level of the flood control system.