For the problems of obscure demand of users, insufficiency and difficulty in correct acquisition in the process of knowledge retrieval, this paper combines with basic theory of interactive genetic algorithm to discuss its application model in the process of requirement and acquisition of knowledge retrieval. And then, it uses the interactive information produced in the retrieval process to make improvement on matching method based on concept map. It uses the interactive information produced in the retrieval process to calculate weight of users to search concept, which enhances the knowledge matching efficiency of concept map. Through test, it compares that intelligent knowledge retrieval system based on ontology has double advantages with high intelligence and efficiency.