The evolution of the Internet and the number of users require the establishment of more sophisticated and efficient systems that must keep track of the entourage. This development requires the establishment of fast and sharp and specialized IRS, to answer a very specific need. In response, these systems carry out several steps, including the enrichment or reformulation of queries. The enrichment of queries consists to add terms to the initial query to find the right document. However, the problem of non relevance of results returned by the information retrieval systems is always present. A simple user seeking accurate information within a specific time is lost before a huge mass of information related to his field of research but who do not necessarily represent his need. Several research studies were launched to address this problem by providing both enrichment approaches, indexing, … but without resolving it efficiently. This work falls within the same goal, it helps to enrich the user request using a domain ontology by adding terms (from ontology) to the user query to target the area research and return a relevant result. The presented approach will be integrated into the IRS of a meta search engine dedicated to a specific domain. The ontology used is composed of a set of concepts linked by semantic relationships, the idea is to extract from the set of ontology concepts (as hierarchy), which represent the field searched by the user. This approach is based initially on the interaction with the user to specify his needs and then, with our approach, on selecting appropriate ontology concepts.