The Land Use Scanner (LUS) is a decision support system for land use planning that is widely used in the Netherlands. The model yields solutions that have a high resolution (500×500 m) and reflect economic equilibrium in the land market, in which eighteen types of land use and nine aggregated habitat types competitively bid for locations. Here, an extension to the LUS is presented that enables the internalisation of environmental externalities. The focus in this study is on two important externalities, namely ammonia emissions from agriculture and habitat fragmentation. We simulate four IPCC-based land use scenarios and assess the level of fragmentation of the nine habitat types. We also perform sensitivity analyses in the most market-oriented scenario on the effectiveness of market-based instruments in promoting more efficient land allocation. The simulation results suggest that particularly the habitat type 'rain-fed mires and bogs' is susceptible to fragmentation. The sensitivity analyses show that internalising the damage costs of fragmentation resulted in smaller changes in land allocation than internalising those of nitrogen deposition. For both instruments, the policy effort was large in relation to the efficiency gains in land allocation. © 2007 Elsevier Ltd. All rights reserved.