Classifying biomedical literature becomes one of the important and challenging tasks lately, due to the fact that a large number of biomedical articles are divided into quite a few subgroups in a hierarchy. It causes difficulties for the researcher to effectively and efficiently organize and retrieve relevant information from the database. In the past, most approaches used in text classification task have applied flat classifiers that ignore the hierarchical structure and treat each concept separately. Therefore, this article presents an exploration of the application of hierarchical structure for classifying huge collection of biomedical text abstracts downloaded from Medline database. To accomplish this goal we will construct, as an example, the disease hierarchical structure with some simple relations from biomedical text abstracts. Then, we will enrich the learning ontology before adapting it to ontology alignment process for classification purpose. Finally, we will evaluate the effects of using hierarchical structure by comparing the accuracy with previous approaches in classification performance.