P2P (Peer-to-Peer) based service computing has emerged as an important new field in the distributed computing arena. It focuses on intensive service sharing, innovative applications and compositions, and in some cases, high performance orientation. However, the main challenge for P2P-based service composition process is how to intelligently mining and selection the most appropriate peers to execute the service application in complex and dynamic situations. Traditional methodologies are still very inadequate to effectively and autonomously conduct the service mining and selection in a real-time environment, as they seldom consider and focus on dealing with complex situations, such as simultaneously considering peers' multiple specifications which reflect different properties of e-services. Different ontology based e-service profiles have been proposed to enhance service oriented framework for the total or partial automation of service mining, selection and composition, which are involved in either centralised or decentralised deployment of services. In this chapter, we propose a modelling based approach to design and develop a P2P based service coordination system and their components. The peer profiles are described with the WSMO (Web Service Modelling Ontology) standard, mainly for quality of service and geographic features of the e-services, which would be invoked by various peers. To fully explore the usability of service categorisation and mining, we implemented an ontology driven unified algorithm to select the most appropriate peers. The UOW-SWS prototype also shows that the enhanced peer coordination is more adaptive and effective in dynamic business processes.