The Component Library System needs to solve two core problems,which are component classification representation and component retrieval. Meanwhile, to implement component efficient retrieval is based on the component rational classification. Aiming at the subjective factors of the faceted classification representation, the component is described by the way which combines the text retrieval with the faceted classification representation, and according to the user requirements, different ranks are given to the main facets of the faceted classification schema. Then, a component clustering method based ranking strategy is proposed. The experiments prove the method is effective, which  implements a more rational classification of the components, and improves the efficiency and accuracy of component retrieval.