Abstract: With the rapid development of information science and communication technology,communication has generated and accumulated a large amount of data and information. In recent years, the development of communication science is rapid with abundant research direction and subjects, but its subject development context and hot research topics urgently require systematic and intuitive analysis and display. In order to help explore and reveal the hot topics and popular subject-method in this field, this paper proposed LDA ( Latent Dirichlet Allocation) topic extraction method based on the perspective of classification. Taking the the periodical literature on communication field as the research objects, it used LDA topic model to extract the topic of classified literature collections and divided the selected 30 hot topics into two categories: subjects and methods. Through the deep analysis of subject-method, it revealed the knowledge point of hot topics. Experimental results indicated that this method can detect discipline theme and research hot spots in a more comprehensive and in-depth way.