Landscapes are complex systems. Their dynamics are the result of multiple interacting biophysical and socio-economic processes that are linked across a range of spatial, temporal and organizational scales. This cross-scale linkage and coupling of radically different subsystems gives rise to an inherently uncertain and self-organized dynamics, characterized by emergent structures and patterns in space and time. Understanding and describing this type of dynamics poses enormous challenges and demands the use of new multi-scale approaches to modelling and analysis. Experience in modelling complex systems shows that a "bottom-up" approach may be the most appropriate. In this approach, the basic modelling units are individual entities at a low level of the system hierarchy (e.g., individual stakeholders, individual animals or plants, or highly resolved landscape units) and higher-level entities are allowed to emerge as the result of the collective actions of the lower-level entities. Such models can be used to simulate emergent system responses to different policy and management scenarios, increasing our understanding of important drivers and feedback loops within the landscape. We use two Canadian examples to show how this approach can be used to inform and support decision-making for landscape-scale, integrated resource management.