Abstract: Conceptual diagrams play a pivotal role in basin flood control efforts, offering visual representations of basin water systems and the topological relationships of engineering facilities. They are indispensable in flood forecasting, engineering operation, and planning. However, traditional manual conceptual diagram methods are inefficient, error-prone, and slow to update. This paper proposes a dynamic approach to constructing digital conceptual diagram, which is based on the extraction of river water system and other flood control elements from multi-source data. The dynamic approach automatically constructs topological and confluence relationships, and generates multi-layer vector digital conceptual diagrams using spatial generalization techniques. This automated process enhances efficiency and accuracy in digital conceptual diagrams, create a visual digital abstraction diagram representation, enabling dynamic updates. Ultimately, this approach furnishes effective tools for scientific and precise flood control measures.