Using agent development tools to construct an agent-based system is a well applied approach.However,the development tools usually do not have the function to check the feasibility about the workflow of the agent system during it implementation stage.Therefore,to develop an evaluation approach to analyze the feasibility of a developing agent system such that the improper workflow of an agent system can be found in the early design stage is a necessary task to reduce the risk of implementation.In this research,a Petri Net(PN) based three-stage evaluation approach was developed.In the conceptual stage,the pitfall of the current agent system developing process was examined and an improvement analysis process was specified.Then,in the system design stage,an evaluation approach which extracted the process log file from a developing agent system into a PN model in terms of a process mining approach-α algorithm was proposed.This model was simulated in a PN simulation package.The agent system performance was evaluated in terms of analyzing the deadlock phenomena of the PN model.Finally,in the implementation stage,the proposed concept was implemented by using an agent developing tool JADE and a PN simulation tool CPN.An agent-based robotic assembly system was used to examine the possible deadlock of the agent system. Using agent development tools to construct an agent-based system is a well applied approach.However, the development tools usually do not have the function to check the feasibility about the workflow of the agent system during it implementation stage.Therefore, to develop an evaluation approach to analyze the feasibility of a developing agent system such that the improper workflow of an agent system can be found in the early design stage is a necessary task to reduce the risk of implementation. In this research, a Petri Net (PN) based three -stage evaluation approach was developed. In the conceptual stage, the pitfall of the current agent system developing process was examined and an improvement analysis process was specified. Then, in the system design stage, an evaluation approach which extracted the process log file from a developing agent system into a PN model in terms of a process mining approach-α algorithm was proposed. This model was simulated in a PN simulation package. The agent system performance was evaluated in terms of analyzing the deadlock phenomena of the PN model. Finally, in the implementation stage, the proposed concept was implemented by using an agent developing tool JADE and a PN simulation tool CPN. Agent-based robotic assembly system was used to examine the possible deadlock of the agent system.