Integrated watershed management implies a collective management of the land reconciling ecological dynamics and social processes to ensure a viable and equitable use of renewable resources and to mitigate conflicts. Based on the integration of existing knowledge from different sources and disciplines, this chapter describes the construction of a spatially explicit multi-agent model to analyze the poorly understood interaction between the risk of land degradation and crop diversification and agricultural commercialization of heterogeneous household-based farming systems in a highland Akha village catchment of up-per northern Thailand. In this region, cash cropping on sloping land is commonly blamed by lowlanders for aggravating land degradation. But on-farm agronomic surveys led to the hypothesis that this interaction is far more complex and could be further examined by using an integrative model to explore simultaneously the interaction between the agronomic and socioeconomic components of the system. The simulated behavior of the different model entities is based on previous field observations and measurements. This agronomic simulator represents actual farmers' cropping practices at the field level under different slope and climatic conditions. The social dynamics are taken into account through the representation of three main types of households identified through a farm survey. They have contrasting historical backgrounds, are managing different amounts of resources, and correspond to a gradient of integration into commercial agriculture. Because key agro-ecological and socio-economic processes need to be simulated at different pertinent scales, this multi-agent model is loosely linked to a geographic information system (GIS) displaying the distribution of three complementary spatial entities in the catchment. Following a presentation of the selected integrative modeling approach, the model conceptualization, its architecture, and modeling sequences are describe