While science has long been regarded as the sole driver for development, it is nowadays acknowledged that inclusion of stakeholders and different scientific perspectives is required to improve our understanding on socio-ecological systems (SESs). To integrate various perceptions and knowledge, there is a growing interest for Agent-Based Models (ABMs) as participatory tools for understanding SESs and their likely evolutions. Because they are centered on individuals, ABMs enable the user of a simulation to visualize his system in a tangible way. Consequently, he can project himself in the simulated word by assuming the role of an agent. Because he can also observe the system at the global level, he can better understand the consequences of decisions and actions. Some experiments using ABMs combine autonomous processes with agents' actions decided by the actors. Such “hybrid agent” simulations enable the stakeholders to interact with the model by modifying the behavior of the agents and the way they use the resources. Therefore, we can collectively explore scenarios to better understand how a desired situation may be reached. This may feed back into the collective design of the model. This talk will include two parts: (1) a brief overview of the ABMs and the pros and cons to use them with stakeholders as a participatory modeling approach and (2) an example of experiment with local communities in an Amazon floodplain in Brazil coping with global changes. A hands-on demonstration of this experiment will be available to the workshop participants to play with the model to better understand the advantages and the limits of such methodology. Beyond the technical and organizational considerations for designing and using such tools, we address the issue of power plays and of how to integrate in the process all the involved stakeholders while strengthening the capacities of the most vulnerable. (Texte integral)