Component library technology is significant direction of national and overseas software engineering research domain. Many national schools and enterprises are getting realized the importance of software component technology. Component library is the component reuse base establishment which supports mass components' description, store, management and retrieve. With the mature of the component library related technology, the research institutes and software enterprises of china have built up series of component library which have significant practice value. However, we are still at the elementary phase of this research. This dissertation is focused on the faceted classification in terms of combining the specialties of multi-language component. Moreover, in order to support the multi-library retrieval, we bring forward the unified faceted classification uniting algorithm. On the basis of specialty of classification schema of multi-library, we improved the unified faceted classification uniting algorithm, which decreased the complexity of the algorithm. The retrieval strategy is on the basis of faceted classification. The terminology of the faceted is leveled, nested vocabulary, considered of which, we use tree matching algorithm as the kernel technology of retrieval. We set an extra level described string to describe the component in order to increase the efficiency of retrieval. At last, we design main model of multi-library retrieval. The goal of our subject is to design a method to solving the faced classification problem of multi-language component, moreover, to support multi-library retrieval. We improved two algorithms: unified faceted uniting algorithm and tree matching algorithm to help the user do search components in different component library, which greatly improved the efficiency of retrieval, finally to reach the goal of improving sharing component resources.