Learning partners are important resources in opening virtual studying community. The tutors who come from the learning partners can help the general learner out of learning disabilities, and improve their ability of communication each other. After constructing a Knowledge Database based on ontology, we record the learner characters about interest, grade of cognize, degree of enthusiasm. According to these factors, we bring forward a learning partner recommendation algorithm using Thematic Study. The result of experiment indicates that the algorithm can make different partner grade on the foundation of the learner's study experience. These grade can reflect the reality situation. It is significant for opening virtual studying community to improve its personalization and intellectualization, consequently to advance study efficiency.