Counter-acting forces to increase rural production and/or its efficiency, and to sustain an ecosystem now recognised to be under increasing and destructive pressures have created exigencies in achieving balanced natural resource management (NRM). Nowhere is this more apparent than in the management of the Murray Darling System in south-eastern Australia. All actions affecting natural resources impinge on the ecosystems that support those resources, the economy based on them, and the human society and culture connected to them. Change is best managed with the cooperation of those most affected. If NRM is to be achieved through informed community decisions, there is a need for a multidisciplinary process, drawing on specialist (intra-disciplinary) expertise, and a requirement to pull the resultant knowledge into an integrated form which supports decision-making at the management and community level. We propose a framework that identifies tasks necessary to support community decision-making and inject specialist technical knowledge into the process. For complex NRM issues, it is likely that there is insufficient information in one or more disciplines to support a strong decision. Where possible, this should lead to the interposition of targeted pilot trials, based on principles of adaptive management, prior to the final assessment and (presumedly) management plans. These 'management experiments' follow a similar path to specialist hypotheses and measurements (based on the same management intervention) followed by an integrated assessment. It appears that identification of, and engagement with, components of the community, and analytical techniques to support integrated assessment are two major areas in which new knowledge is urgently needed.