<p>Semi-natural grasslands and wooded pastures are biodiversity-rich, declining habitats whose conservation depends largely on voluntary management by farmers. Their management is supported through agri-environment schemes. Effective scheme design benefits from trans- and interdisciplinary approaches integrating ecological goals with social constraints and human behaviour. This study tested a revealed-behaviour approach by modelling relative agri-environment contract feasibility across Finland using Maximum Entropy (Maxent) based on realised contract uptake and landscape predictors. The models reproduced spatial patterns of contract locations reflecting historical habitat distribution, land-use change, eligibility criteria, and joint decisions by landowners and farmers. Spatial associations and response curves identified structural constraints and opportunity patterns influencing uptake, providing a spatial layer to support ecological prioritisation and scheme design. Maintaining farm types capable of integrating grazing or mowing into routine operations appears critical where semi-natural habitats remain, while payment design should better reflect variation in management costs among farm types and sites.</p>