This research is dedicated to the improvement of medical information retrieval (IR) systems, specifically in the extraction of more relevant and credible documents. By harnessing ontologies and machine learning (ML) techniques, our approach incorporates key elements such as modular ontology building and a specialized relevance ranking algorithm. We also integrate machine learning and information retrieval to offer a deeper analysis of semantics of natural language used in the query and medical documents. Our proposed approach supports users in searching for medical information on the Web by offering a semantic layer that helps enriching their query using ontology building as well as assessing the credibility of returned documents for trustworthy results. The evaluation results underscore improvement in the relevance of returned documents as well as their credibility which offers not only pertinent information but also trustworthy documents that the user can rely on.