Abstract: Ontology is widely used in semantic web to model the knowledge in the form of classes or individuals and relationships among them. Ontology languages such as Web Ontology Language (OWL) and Resource Description Framework (RDF) are built based on discrete logic which cannot deal with probabilistic knowledge about a domain. Various approaches have been made to represent uncertainty in ontology, such as Bayes OWL, Multi-Entity Bayesian Networks (MEBN) and Probabilistic OWL (PR-OWL). These research emphasize on how to represent uncertainty and reasoning in ontology based on Bayesian network approach. On the other hand, sometimes we need to solve a problem with the logic and probabilistic approach simultaneously as well. This study discusses the importance of combining both approaches. The research also proposes an approach or framework to perform both logic and probabilistic reasoning that can be implemented into an ontology-based application. A prototype has been developed as an experimental in order to simulate the work of this framework by having two cases: investor problem and social CRM for higher education.