Natural resource management (NRM) typically involves complex decisions that affect a variety of stakeholder values. Efficient NRM, which achieves the greatest net environmental, social and financial benefits, needs to integrate the assessment of environmental impacts with the costs and benefits of investment. Integrated assessment (IA) is one approach that incorporates the several dimensions of catchment NRM, by considering multiple issues and knowledge from various disciplines and stakeholders. Despite the need for IA, there are few studies that integrate biophysical modelling tools with economic valuation. In this paper, we demonstrate how economic non-market valuation tools can be used to support an IA of catchment NRM changes. We develop a Bayesian Network model that integrates: a process-based water quality model; ecological assessments of native riparian vegetation; estimates of management costs; and non-market (intangible) values of changes in riparian vegetation. This modelling approach illustrates how information from different sources can be integrated in one framework to evaluate the environmental and economic impacts of NRM actions. It also shows the uncertainties associated with the estimated welfare effects. By estimating the marginal social costs and benefits, a cost-benefit analysis of alternative management intervention can be gained and provides more economic rationality to NRM decisions.