Automated techniques and tools are required to effectively locate services that fulfil a given user request. To this purpose, the use of semantic descriptions of services has been widely motivated and recommended for automated service discovery under highly dynamic and contextdependent requirements in distributed environments. Our aim in this work is to propose a semantic-enriched framework to describe services and an ontology-based hybrid approach where such framework is exploited combining together different kinds of comparison strategies to provide a flexible and efficient matchmaking between service descriptions. For service discovery two matchmaking strategies are proposed: a deductive strategy based on Description Logics, with a reasoning procedure exploiting ontology knowledge to assess the type of match between services; a similarity-based strategy, exploiting retrieval metrics to measure the degree of match between services.