Aiming at the problems of complex display elements, strong timeliness of tasks, difficulty in capturing key information, slow transmission of key instructions, and high load of user interface in the situation information system, this paper studies the adaptive method of the situation information system interface. The situational information is divided into three categories: system data, system events and user interaction behavior. Firstly, the data is preprocessed, spatiotemporal aligned and other operations are carried out to realize multi-source data fusion. Then, the context awareness model is established through BERT semantic understanding model, multi head attention mechanism and reinforcement learning. Then, an adaptive interface generation mechanism that can transform situational information into an interface is established. Finally, the intelligent push of interface content and adaptive presentation of interactive components are realized.