ABSTRACTOpen source product design is characterized by dynamic evolution, autonomous task control, loose organization structure, and frequent collaboration, which shows the process is full of high uncertainty. Firms have used this innovation pattern to obtain ideas. To describe the firm-leading self-organization evolution process, an agent-based modeling and simulation method is applied. We have proposed an agent-based model including designers and firms in light of Complex Adaptive System theory firstly; Secondly, their behaviors, especially the two typical behaviors: task selection and plan screening are illustrated; Finally, the simulation results show that:1) evolution process of agent, product, and community can be described clearly by simulation. Agent-based simulation can support the analysis, evaluation and optimization of the evolution process;2) in task selection phase: firm’s recommendation can improve the evolution efficiency;3)in plan screening phase, firms’ “control” also plays important role in the evolution process, but the impact of designer agents’ self-organization cannot be ignored.