The Semantic Web Ontology learning technology introduced is to reduce the overall time of ontology construction. Ontology construction process includes several aspects, and extraction domain concept is one of the most important aspects. In this paper, we proposed a domain concept extraction algorithm. The algorithm aims at improving the extraction efficiency and accuracy of domain concepts, thereby enhancing the efficiency of building the entire ontology. We have achieved integrating personalized recommendation technology and concept extraction technology into a coherent whole, and more accurately and efficiently finished extracting domain concepts. Through the analysis of the results gained from the experiment proved the feasibility of the algorithm.