Traditional natural resource management policy has largely focused on implementing prescriptive solutions to maximize a production function. The fundamental assumptions of this approach were: (1) that ecosystems behaved in a linear, deterministic manner; (2) that there was general community agreement on the value of different ecosystem services; and (3) that land managers would accept and adopt the recommended technology. The result has generally been an unpredictable performance by ecosystems, conflicting expectations among users, and low adoption rates for the outputs of research and development (R&D). We propose that an approach that integrates the fundamentals of nonequilibrium ecology and "soft" systems methodologies to define options, make management decision recommendations, and implement programs will result in improved predictability of ecosystem response, more realistic expectations on the part of users of ecosystem services, and better uptake of technology by land managers.