Abstract: 【Purpose/significance】By clarifying the public intelligence needs, it points out the direction for the development of smart intelligence service research on major emergencies, which helps the collaborative solution of complex problems, and then improves the overall emergency response level of the society.【Method/process】Taking the 2021 Henan rainstorm event as an example, based on the event mapping method, combining the LDA model and the UIE model, through the steps of theme extraction, entity extraction and relationship extraction, the potential knowledge structure under the event theme is revealed, and the evolution logic of the internal rela⁃ tionship of the public intelligence needs is clarified. The whole perception process can consider the subjectivity of public perception and the objectivity of automated data processing.【Result/conclusion】Intelligence needs exhibit characteristics such as continuity, fo⁃ cus, correlation, secondary generation, derivativity, and evolution. Intelligence that serves public perception needs to have the function of guiding public behavior and decision-making, explaining the causes and consequences of events, and satisfying the public´s emo⁃ tional needs.【Innovation/limitation】The internal logic of the public´s intelligence needs has been clarified using knowledge graphs, and the analysis results will be further improved by optimizing the technical model in the future.