Abstract: The world wide web is a vast resource of information and services that continues to grow rapidly. Developing an automatic classifier, which has ability of classifying documents into appropriate categories predefined in the topic structure based on document contents is a crucial task. Traditional methods of documents classification need characteristic abstraction and classifier training. The work of collecting trainable text terms is laborious and time-consuming. In order to solve the problem, this study proposes an ontology based approach to improve the efficiency and effectiveness of Chinese web documents classification and retrieval. First, the approach establishes an ontology model based on knowledge base. Second, it creates ontology for each subclass of the classification system. It uses RDFS to convert knowledge into ontology and to define the relations among ontology. Finally, web documents classification is performed automatically using the ontology relevance calculating algorithm. Present experiments show that the accuracy of ontology based approach is very close to most classical methods includes Support Vector Machines, K-Nearest Neighbor and Latent Semantic Analysis. Additionally, ontology based algorithm is more stable and robust and can obtain better recalling rate than other three methods.