Uncertainty about the link between policy and physical outcomes and the valuation of differences in outcomes is pervasive in environmental policy. This paper develops and refines methods for improved policy analysis in the context of environmental-policy areas that emphasize each type of uncertainty. Outcome uncertainty is illustrated by the possible link from anthropogenic chemical emissions to stratospheric ozone depletion and to the ultimate consequences in the biosphere. The author develops methods to characterize and assess uncertainty about these links, and applies them to describe likely future emissions in the absence of additional regulations. He then develops a decision framework that recognizes expected learning about these links, and applies it to provide specific guidance for current policy. He discusses value uncertainty in the context of valuing reductions in human health risks, specifically mortality risk from residual pesticides in food. He arrives at preliminary estimates of consumer valuation of pesticide-free produce and the corresponding risk reduction using each of the leading classes of methods — revealed preference and contingent valuation.