Abstract: The identification and analysis the livelihood topic in forum is an important research direction.As a realtime public information platform,forum topics are divided into sudden topics and persistent ones which are close to the life of the people's livelihood,and our research mainly focuses on the analysis of livelihood topics.Identifying livelihood topic mainly combines the LDA model and similarity model on the timeline,through the correlation analysis between the adjacent time slice,gets the persistent livelihood topic which has time slices more than threshold periods.The methods in real data sets validate the effectiveness.