Ideal information services should be customizable, personalized, and just-in-time assistive at the point of need. In this paper, we describe an Internet-based information retrieval scheme to deliver assistive information services, meeting users' dynamic and changing needs. The proposed approach focuses on the effectiveness of information aggregation and personalization. Domain knowledge and users interest models are essential in this research. Using topic-oriented search services to improve the effectiveness of retrieving domain knowledge from the Internet, while using long-term and short-term user profiles meets the need of being assistive. An intelligent service system based on RSS technology is built, aimed at validating the applicability of the proposed methodology.