We describe a landscape scale non-linear discrete choice spatial optimisation model for identifying cost-effective strategies for achieving environmental goals. Spatial heterogeneity and configuration issues such as fencing costs, patch sizes and network linkages are explicitly accounted for and quasi-optimal allocations are determined using simulated annealing. Applications of the model being developed with New South Wales Catchment Management Authorities are discussed. These focus on targeting investments in revegetation to control dryland salinity and erosion and provide biodiversity benefits whilst minimising direct and opportunity costs. We compare our approach with alternate investment approaches.