Abstract: Most of the existing peer-to-peer (P2P) systems suffer from the lack of the support for semantic contents, thus making them unsuitable for knowledge sharing. To solve the problem, a prospective method is suggested, introducing ontology to describe the resources to be shared in P2P systems semantically. In a pure P2P environment, where the different peers are keeping up their characteristics of locally ontological heterogeneity individually, an HOM (history-based ontology matching) method is proposed on the basis of historical query logs. Mining the possible mapping relationship between different ontological concepts from historical query logs, the proposed method applies VSM to the recognition of the possible linguistic relationship between ontological concepts. Then, the mapping relationship between concepts is corrected to match the ontological concepts between different peers. Analytic and experimental results show that, compared with the existing methods, the HOM method considers the autonomy of peers and is more precise on matching different ontology.