Changes in the structure of agriculture are known to affect emissions of environmental pollutants from agriculture. Such changes are often driven by structural changes in agricultural production, so structural changes are likely to have indirect effects on emissions. In a pilot study, we consider how linking two complementary simulation models might be used to explore these effects. The agent-based AgriPoliS model was used to simulate the structural dynamics of agricultural production. The results from AgriPoliS were passed via a number of intermediate models to the Farm-N model, which was used to estimate the nitrogen surplus and losses from each farm for each year. The modelling complex was exercised by simulating the effects of two plausible policy scenarios for each of 14Â years. The initial sizes and types of farms were based on statistics from a region in Denmark and the farms were randomly distributed within this region. The reference scenario (REF) implemented the current area-based Common Agricultural Policy payments for Denmark. The 1Â LU scenario applied the additional constraint that a minimum area of 1Â ha land had to be available for the application of the manure produced by one livestock unit. Substantial changes in the structure of agricultural production were shown for both scenarios. The effect on the regional nitrogen surpluses was predicted to differ between scenarios and the contribution of the different farm types to change with time. Predicted ammonia emission changed with time and differed between the scenarios, whereas the Danish fertiliser and manure legislation meant that nitrate leaching remained fairly stable. The implementation of additional environmental legislation significantly changed the trajectory of structural adjustment processes. Results emphasize the complex interplay between structural changes, losses of nitrogen, and environmental regulation. It is concluded that the effects of structural change on environmental emissions can be usefully explored by linking agent-based models of farmers' investment decisions with other models describing nutrient losses from the farm and that such modelling can play a useful role in designing effective environmental policies for agriculture. However, the approach demands the availability or collection of many region-specific data and this could create a barrier to its use.