The field of the web service orchestration introduced to generate a valuable service by reusing single services. Recently, it suggests rule-based searching and composition by the AI (Artificial Intelligence) instead of simple searching or orchestration based on the IOPE(Input, Output, Precondition, Effect) to implement the Semantic web as the web service of the next generation. It introduce a AOP programming paradigm from existing object-oriented programming paradigm for more efficient modularization of software. In this paper, we design a dynamic web service orchestration and invocation scheme applying Aspect-Oriented Programming (AOP) and Reflection for Semantic web. The proposed scheme makes use of the Reflection technique to gather dynamically meta data and generates byte code by AOP to compose dynamically web services. As well as, our scheme shows how to execute composed web services through dynamic proxy objects generated by the Reflection. For performance evaluation of the proposed scheme, we experiment on search performance of composed web services with respect to business logic layer and user view layer.