[Purpose/significance] Currently,the emergency decision-making of unexpected events is a major demand of the country,which should be improved to the level of emergency decision,adapt to rapid response requirements of mass information,high precision,and cross disciplines,as well as provide emergency solutions orderly,efficiently,and timely under the structure of knowledge base collaboration. [Method/process]Firstly,using the knowledge organization theory and methods,this paper uses deep knowledge discovery from scenario evolution and mapping rules of knowledge base to study the knowledge representation oriented to problem features in scenario evolution,and to describe the semantic knowledge acquisition of the scenario. The collaborative architecture of knowledge base for emergency decision-making is proposed. Secondly,the paper constructs the feature dictionary for problems from both the methods of vocabulary classification and the standards of classification and staging for various emergencies. Finally,the paper carries out thorough research on knowledge base collaboration based on scenario deduction,and scenario-driven knowledge reasoning and calculation. [Result/conclusion] The problem that a single knowledge base cannot overcome the multi-field and evolution problems of emergency decision is solved.