Semantic annotation has gained momentum with the emergence of the current user-generated content paradigm on the Web. The ever-growing quantity of collaborative data sources has resulted in the requirement for efficient approaches to create, integrate and retrieve information more efficiently in an environment where the users ask for accurate information. The main research challenge of the current work is using manual semantic annotation in a highly accurate closed domain, a conceptual domain with a minimal set of concepts where the benefits of adding semantics, search efficiency, optimization and the cost estimations are viable. This paper presents a semantic annotation approach for highly accurate closed domain based on multi-ontology annotation (domain and application ontologies).