Component software is one of the most promising ways for the software reuse. The tremendous increase in the use of World Wide Web (WWW) has accelerated opening reuse repositories on the Internet. However, the existing document search engines are not proper for the software components on the WWW, because they usually conduct HTML-based keyword search. In this paper, we describe the intelligent search agent system for components on the WWW. The agent has intelligence such as preference, reasoning, and learning. In order to improve the search performance and user's convenience, we adopted several progressive technologies and XML with rich semantics.