One important aspect in self-directed learning is to find learning resources, which can be referenced from the Internet. In the process of finding the resources, users have experience in accessing certain sites for specific information. But, when they want to refer back to related information, usersare supposed to recall the domain of sites and to determine the proportion of priority on a number of sites. The experience of accessing the Internet is mapped into semantic representation of Knowledge Organization (KO) like a memory thatit can be exploited to support in gaining new resources in the same domain of knowledge. This study aims to develop a web retrieval application to assist users based on their experience. The application will identify the domains of the sites that were previously accessed by tracing the representation. The domains are used as references to direct the search. Searching the web resources is undertaken by the composition of domain based search and search without domain limitation. The composition ofboth searchescan be determined by the users according to their expectations. The domain based search composition is determined by the resourcecontributions within the scope of a knowledge domain. The test results show that the system is able to identify the domain’s location and the proportions as preferences for searching. The performance of the system in searching shows thatthe query returnsof relevant documents are dominant.