In this paper we are concerned with optimal investment decisions when dealing with land allocation problems. We aim to emphasise the importance of flexible modelling in order to capture irreversibility. In particular, we stretch a discrete model, firstly developed in Coggins and Ramezani (Coggins & Ramezzani , 1998), in order to cover a more realistic and complex scenario. Both environmental and economic uncertainty are included in the model and treated using an integrated approach, in which decision analysis techniques and option pricing theory are jointly applied to evaluate development versus conservation opportunities. Moreover, we take explicit account of how uncertainty interacts with two types of irreversibility: sunk costs associated with investment in developing decisions, including environmental and social costs due to environmental degradation, as well as sunk costs associated to environmental regulation and conservation. Finally, we use the Quasi Option Value, to derive decision rules that account for different levels of flexibility of land allocation possibilities.