The process of document annotation begins with the syntactic process of the unstructured document which we focused on the academic theses. The basic linguistic process of tokenization, sentence splitting and lemmatizing is done and the term weight and frequency is calculated. The structured terms which are stored in a normal database will be map to the domain ontology. For our research study, we used the ACM topic hierarchy which is a lightweight domain ontology. In order to support semantic search, each lemmatized term stored in normal database will be matched to the related concept in the ontology using label presented in the ontology instances. If a match is found, the concept URTs is added to the Annotation Class. For example, refer to figure 2, the lemmatized term of "Arifah Alhadi" will be notified as a label and matched to the labels presented in the thesis.owl. Once the match is found, the annotation is created between the term and the document. The URIs of the instance and the related concept will be added to the Annotation Class. The instance of "Arifah Alhadi" is a "Studentl" under the concept of "Student" which is a subClassOf "Creator" and "Person". All the inferred class will be annotated and stored in the Annotation Class. The inferred class to the instance "Studentl" will be "Student", "Creator" and "Person".