The topic text and the semantic network are used to mine the comment texts in the travel e-commerce, thereby guiding consumers and businesses to make important decisions on the comment information. This paper proposes a method based on LDA (Latent Dirichlet Allocation,LDA) topic clustering and semantic network model (LDA topic clustering and semantic network model，LTC-SNM) to study the online commentary text of hotels.Firstly, the online review text is obtained for data preprocessing, Word2vec is used to generate the word vector, and the machine learning algorithm is used to classify the comment text. Secondly, the classified text is clustered by the LDA theme model to generate the hotel’s feature keywords. Finally, through ROSTCM, feature subject words and modified emotional words are generated into a semantic network, which alleviates the complexity of mining text information. The experimental results show that the proposed LTC-SNM text mining method makes the topic of online user evaluation more expressive.