[Purpose/significance] In recent years, it has become a trend to introduce big data technology to integrate and manage online public opinion. How to accurately identify the development stage of public opinion is an important content of public opinion governance. [Method/process] Based on the network public opinion system theory and the network public opinion life cycle theory, the article focuses on the decline period of the network public opinion in emer? gencies, and builds a research and judgment index system for the decline period of the network public opinion in emer? gencies; the data is obtained through Python crawler technology, combined with SKEP Algorithm and RBF neural net? work for multi-case empirical analysis. [Result/conclusion] This paper constructs an index system for judging the de? cline period of network public opinion in emergencies, and the RBF neural network verifies the feasibility of the index system, and proposes a guiding strategy to deal with the decline period of network public opinion in different emergen? cies. It provides corresponding theoretical references for government network public opinion governance, locating key nodes of network public opinion, and accelerating the decline of network public opinion.