Sustainable development is a difficult phrase to define, particularly in the context of human ecosystems. Questions have to be asked, such as “Sustainable for whom?” “Sustainable for what purposes?” “Sustainable at the subsistence or at the luxury level?” and “Sustainable under what conditions?” In this paper, development is taken to mean improving the quality of life. (If development were to mean growth , then it could not be sustained over the long term.) Studies of development must, of course, consider economic factors, particularly in the case of societies who suffer from the pollution of poverty. However, cultural and environmental factors are equally important. In fact, development is not sustainable over the long term if it is not ecologically sustainable. The terms maximum sustainable yield of a renewable resource, carrying capacity of a region and assimilative capacity of a watershed or airshed are discussed. Approaches using these resource management tools are recommended when external conditions are not changing very much. The problem today is that unprecedented rates of change are expected in the next century, not only of environmental conditions such as climate but also of socioeconomic conditions such as renewable resource consumption and populations (of both people and of automobiles)! In rapidly changing situations, policies must be adopted that strengthen resilence and ecosystem integrity ; that is, society must increase its ability to adapt. Maintaining the status quo is a long-term prescription for disaster. The problem is of course that little is known about how to design strategies that will increase resilience and ecosystem integrity, and this area of research needs to be strengthened. Some suggestions on appropriate indicators of ecosystem integrity are given in the paper but these need considerable refinement. One of the main problems with long-term environmental policy formulation is the uncertainty to be expected, including the possibility that complete surprises will occur, as the time horizon moves forward. Research programmes will of course reduce the uncertainty but will never eliminate it. There is therefore an urgent need to improve current methods of environmental policy formulation, which avoid foreclosing of options and permit continuous review and adaptation of policies. The Canadian round-table approach is mentioned, as well as the policy-exercise approach developed at IIASA (the International Institute for Applied Systems Analysis).