Abstract: Traditional workflow research focuses on process-based workflow property, modeling technology and verification technology. The authors find that there are deficiencies in dealing with the knowledge or data representation capabilities of these process-based workflows. Although workflow can fulfill many requirements for customizing process design and execution, researchers have yet to reach a common view on how to model the dataflow conveniently and legibly, as well as how to verify the correctness of these data-sensitive work flows efficiently. In this work, a novel method is introduced to define a workflow with restricted dataflow and a verification methodology is provided to identify format, syntactic, logic and data semantic errors in this workflow. A Workflow-based Knowledge Service (WKS) system is also constructed to provide a graphic design tool and executing engine of restricted dataflow based workflow. Experiments on WKS demonstrate the capability of this method to model knowledge applications to executable workflows as well as the efficiency of its verification algorithm.