Modern agriculture faces the dual challenges of feeding a growing human population, while preserving natural resources and slowing current trends in climate change and its impacts. A deep understanding of the functioning of agricultural landscapes appears crucial if we are to move towards sustainable, complex and resilient agroecosystems. Modelling is a powerful tool to address these issues since it can inform these transformations by simulating the multiscale ecological flows and myriad interactions agroecosystems host, and the multilevel stakeholder actions and their feedbacks in landscapes. This chapter shows that models can provide guidance on the transition towards future multifunctional agricultural landscapes. We have focused on process-based models, which allow for a more thorough understanding of the underlying mechanisms and how these may be manipulated. We first examine how models can simulate the structure and the dynamics of agricultural landscapes, emphasizing the complex mosaic of urban, peri-urban, rural and seminatural habitats. Then, we consider the simulation of biotic and abiotic flows and their complex interactions in the mosaic habitats. Using formalisms from the social sciences, we integrate human decision-making and actions into the landscape models, thereby encompassing a major component in the landscape transformation process. Finally, we outline some avenues for future research. We have focused on improvements to landscape representation, and have suggested ways to bridge the gap between future landscape conception and manipulation, thus providing operational guidance for the transition towards future agricultural landscapes that achieve our objectives.