Abstract: With the impact of global climate change and human activities, flood disasters has become one of the important natural disasters facing China. In order to improve the accuracy and efficiency of flood disaster early warning, this paper proposes a flood disaster early warning platform based on digital twin. The platform adopts advanced digital twin technology, uses technical means such as big data and artificial intelligence to analyze flood data, simulate and predict the development process of flood disasters, and provide more accurate and reliable early warning information. At the same time, the platform can also explore the laws and characteristics of flood disasters, and provides a scientific basis for the government to formulate flood prevention and mitigation strategies. This study provides new ideas and technical support for the construction and design of flood disaster warning platform, which is helpful to improve the flood prevention and response ability, and ensure the safety of people's lives and property.