In this paper, we propose a methodology for addressing trust in Semantic Web Services (SWS) - based applications. The aim is to enhance the capability-driven selection provided by current SWS frameworks with the introduction of trust-based selection criteria. We present an ontology - Web Services Trust Ontology (WSTO) -- that models the context of a trust-based interaction and enables the participants to describe semantically their trust requirements and guarantees. WSTO makes use of WSMO as reference ontology for representing Web Services and embodies the problem of finding the most "trusted" Web service as a classification problem. To test our methodology, we implemented a specific module within IRS-III -- a WSMO-based SWS broker -- and deployed a prototype application based on a use case scenario.