In this new century, there is an urgent need to integrate and organize knowledge into suitable frameworks to examine essential problems with the people involved in solving them. Recent advances in computer science, particularly distributed artificial intelligence and multi-agent systems (MAS), are creating a strong interest in using this new knowledge and technologies for various applications to better deal with the increasing complexity of our fast-changing world, particularly for studying interactions between societies and their environment. By emphasizing the importance of interactions and points of view, the MAS way of thinking can facilitate high-level interdisciplinary training and collaborative research among scientists working in ecology and social sciences to examine complex problems in the field of integrated natural resource management (INRM). This paper describes how an ongoing project based on a series of short courses in the field of MAS, social sciences, and INRM at three different universities in Thailand is transferring European expertise and research results to an Asian audience of graduate and post-graduate students and young researchers interested in innovative and action-research-oriented interdisciplinary approaches. The course structure and organization, and its contents, are described and assessed. The course participants are characterized and their opinions are used to evaluate the strengths and weaknesses of this training program. The first sustainable outputs and key preliminary lessons learned from this innovative collective learning experience are presented. In conclusion, we suggest ways to support the emergence of a regional network of "MAS for INRM" practitioners in Southeast Asia to build on the dynamics begun by this project and serve the need for such interdisciplinary training across Southeast Asia.