Abstract: The knowledge platform in digital twin basins needs to support application requirements such as full mapping of physical basins and interactivity of twin basins.A knowledge system architecture of digital twin basins based on knowledge graphs is proposed for the construction of water knowledge graphs, historical scenario models, application rules, and expert experience in the knowledge platform.Among them, the graph model is used to describe the association relationship between entity concepts in business scenarios, the mechanism and pattern behind the evolution of events are described from the perspective of events, and the business-oriented association network between business-events-elements is formed, and finally a panoramic coupling network is constructed.Based on the panoramic coupling network, we further establish a closed-loop application architecture that interacts physical and digital space, and business and support layers.This architecture is applied to the digital twin basin platform, which is event-driven and forms a closed loop for decision-making through data feedback to support intelligent management and decision-making in the water resources business.