With the component technology development and the component library’s scaling up, representing and retrieving components, the two core technologies of component library management have been a focus of research. Based on the present widelyused component faceted classification description method and its features, and from the different requirements and ways of component retrieval used by component reusers, three component retrieval ways based on browser, facet and term, the five corresponding matching models and two matching algorithms are proposed in this paper. The theoretical analysis and experimental results show that these three retrieval methods can meet a variety of requirements of the component reusers while being used in a largescale component library, and the retrieval efficiency is feasible.