The term "search engine” is traditionally used to refer to crawler based search engines, manually maintained directories, and hybrid search engines. However, current search engines do not fully satisfy the users' needs especially in terms of accuracy and specificity of the results. This thesis proposes an intelligent searchagent system on top of the Semantic Web. The Semantic web is an extension of the current web that will enable web resources to become machine understandable through the use of meta-data. Many languages have been developed for representing metadata in the semantic web such as DAML, OIL, and OWL. The process of adding metadata to the original data is called the annotation process. The proposed system consists of five main parts: the Annotator, the Ontology Parser, the Indexer, the Search Agent, and the Data Repository. Two kinds of search are implemented: keyword/concept based and structured search. The keyword/concept based search matches a user’s query terms to concepts while the structured search allows a user to choose the concept that s/he want to search for together with some attributes for this concept. The proposed system was evaluated using the precision measure. The evaluation is based on the comparison of the proposed system with the Lucene desktop application Evaluation results show that the proposed system’s two search models (keyword /concept based and structured search) achieve better results than Lucene..