The purpose of this study was to evaluate the sustainability of agroecosystems. The framework developed within this study is systems-based with the dynamic linkages between the system components explicit. The primary objective of the study was to develop a computer model, the Sustainable Agroecosystem Model (SAM), that dynamically integrates the economic and ecological components of an agroecosystem. The model was used to assess the sustainability of agroecosystems, defined by ecodistrict boundaries, in the Brown soil zone of southwestern Saskatchewan. The SAM was comprised of three components: (1) a soils model that simulated soil and crop growth parameters; (2) an economic model that simulated land use and cropping decisions; and (3) a habitat model that calculated habitat and biodiversity parameters. These components were largely self-standing models comprised of important processes of the soil, economic and ecological sectors of the agroecosystem respectively. To simulate the co-evolutionary changes of the agroecosystem the component models were dynamically linked, based on a one year time step, through selected input and output parameters. The output of the component models reflect elements of the natural and man-made capital stock of the target agroecosystems and were used as sustainability indicators. The concept of strong sustainability was adopted in the analysis such that changes in these indicators signal changes in the relative sustainability of the system. The study focused on two types of simulations: (1) the relative sustainability of four ecodistricts was assessed using baseline simulations. This analysis highlighted the importance of biophysical constraints to the sustainability of an agroecosystem. These simulations indicated that the development of production technologies and policy initiatives, targeting agroecosystem sustainability, should explicitly consider the regional biophysical constraints faced by farms; and (2) the relative sustainability of a single ecodistrict subjected to economic (carbon credit and carbon tax policies) and environmental (climate change) perturbations was evaluated. These simulations highlighted the difficulty in identifying a single policy that leads to a sustainable agroecosystem. In general, policies that resulted in improvement in some components of the capital stock caused degradation of other components. The Identification of preferred policy, in terms of agroecosystem sustainability, requires a weighting of system effects based on societal preferences, ethical responsibilities, degradation thresholds and system co-evolution.