This paper addresses the service selection issue via a service agent framework coupled with a computational model based on trust ontology. The trust ontology proposed from the socio-cognitive view can offer a more insightful understanding in the trust relationships among service agents. A set of trust computational rules have been given to support the trust analysis and reasoning. Also the structure of the service agents has been given. Finally, a case study is used to illustrate how the service agents infer the trustworthiness of their partners by using the set of trust reasoning and computational rules.