Abstract: 【 Purpose/significance】 In order to enhance the response efficiency and credibility of the government in major emergencies, this study aims to construct a new media based public opinion intelligent decision-making intelligence system, which can accurately capture public demands and concerns and assist the government in making scientific decisions. 【Method/process】 First, use Python crawler technology to automatically collect relevant public opinion information from new media platforms such as microblog, blog, We⁃ Chat official account, etc; Then, LDA and K-means algorithms are used to extract and analyze topics, and information is integrated based on topic consistency; Secondly, combining the three stages of public opinion development (growth period, deceleration period, saturation period) with the three levels of subject, theme, and emotion, a public opinion situation analysis model is constructed to moni⁃ tor the dynamics of public opinion; At the same time, establish a public opinion risk warning model, predict public opinion risks through historical data learning, and automatically generate intelligence reports; Finally, taking the "Gansu earthquake" event that oc⁃ curred on December 18, 2023 as an example, the constructed new media intelligent decision-making intelligence system for major emergencies is empirically verified.【 Result/conclusion】 The empirical results show that the system can efficiently and accurately ex⁃ tract keywords, achieve high-quality integration of public opinion information, effectively monitor the development trend of public opinion, and accurately warn of public opinion risks, providing strong support for government decision-making. 【Innovation/limita⁃ tion】 However, the singularity of cases may limit the universality and breadth of research conclusions, and there may be differences in the dissemination and evolution of public opinion for different types of major emergencies. Therefore, future research will introduce more diverse cases to further validate and optimize the constructed system.