Ecosystem management requires the explicit treatment of interactions between humans and ecosystems. An exploratory model for integrating social and ecological dynamics was introduced to study ecosystem management strategies. This paper focuses on the management of lake eutrophication. The model was developed to include the dynamics of the lake, the behaviour of agents using phosphorus for agricultural purposes, and the interactions between ecosystem and farmers. Analyses with the model showed that the dominating type of cognitive processing was a relevant factor in the response to uncertainty and policy measures. A higher target level for returns on the use of phosphorus was found to lead to a more intensive use of phosphorus and to higher levels of phosphorus in the lake. Simulated farmers used phosphorus more intensively in situations with high natural variability. A tax on phosphorus had little effect on the behaviour of the farmers when they felt uncertain and had low target levels for returns. © 2001 Elsevier Science B.V. All rights reserved.