Computing becomes increasing mobile and pervasive today. Building context-aware computing has been drawing much attention from researches. However, it is still a complex and time-consuming task due to lack of appropriate infrastructure or middleware-level support. In this paper, we first propose a formal context-model based on ontology using OWL to realize semantic context representation, context reasoning and knowledge sharing, context classification and so on. Our context model captures various contexts by introducing a classification scheme; captures relationships between different context information by introducing dependency tag to the property associated with a specified context class; captures quality of context by annotating sensed context with extensible quality constraints. It also supports the use of different context reasoning engines to reason about various contexts so that applications can be given a notion of the confidence of different contexts before acting on it. And then, based on the context model, an architecture for building of context-aware services is presented.