[Purpose/significance] Identifying and evaluating online high-quality advices in crisis situations and ob⁃ taining valuable and usable information can help management departments find solutions to the crisis and solve the cri⁃ sis effectively. [Method/process] According to the characteristics of online advice in crisis situations, this study con⁃ structed a Long Short Term Memory(LSTM)-based user suggestion identification model. Based on the Elaboration Like⁃ lihood Model and the Advice Response Theory, this study built a user advice quality assessment system from two dimen⁃ sions of information quality and source quality. The system consists of indicators such as timeliness, relevance, urgency, and so forth. Taking the public event of“the 49th High School Student in Chengdu Falling from a Building on May 9th”, 2021 as an example, this study identified and evaluated the user advices, ranked and analyzed the distribution of quan⁃ tity and content themes. [Result/conclusion] The online user advices in crisis situations can be classified into highquality advices, usable advices and ordinary advices. The quantitative distribution of the three types of advices and the thematic distribution of their content have different characteristics at different stages of the crisis life cycle. According to these characteristics, this study puts for ward suggestions for management departments to apply user advices and opti⁃ mize the crisis response process.