Determining the size, effort and cost of Service-oriented Architecture (SOA) systems is important to facilitate project planning and eventually successful implementation of a software project. SOA approach is one of the recent software architectures that allow integration of applications within and outside the organization regardless of heterogeneous technology over a distributed network. A number of research studies have been done to measure SOA size.However, these studies are not based on software metrics rendering them to be ineffective in their estimation. This study defined a set of SOA size metrics based on Unified Modelling Language (UML). The study employed Briand’s theoretical validation to test the validity of the designed SOA size metrics. Findings from this study will provide metrics to measure SOA size more accurately and form a basis for future software engineering researchers to develop more effective and more accurate size metrics and effort estimation methods.