Farming for food, fibre and other products for human consumption is a dominant land-use throughout the world. Rural landscapes are also critical to the conservation of flora and fauna, and the maintenance of ecological processes on which all of life depends. In Australia, excessive clearing of native vegetation in the most productive agricultural landscapes has had profound environmental and social consequences. Restoration of these landscapes is an enormous challenge that offers the opportunity to shape the future of Australia, environmentally, socially and economically. In this paper we address the issue of identifying priority areas for conservation in agricultural landscapes. The spatial location of conservation actions in rural landscapes is important because it affects the degree of representation of the biota, the level of protection for rare and threatened species, the adequacy of habitats for species and communities and their future viability, the maintenance of ecological processes, and the integrity of habitats. However, because most land in agricultural regions is privately owned, effective implementation of restoration goals in preferred locations requires understanding of social processes, recognition of pragmatic issues in land management and financial commitment by the wider Australian society. We briefly review the strengths and limitations of some current approaches to determining priority locations for conservation action, including the use of general principles, species-based approaches, quantitative approaches for assessing representativeness, and "bottom-up" approaches based on landholder action. There is no single "best" solution: the most effective approach or combination of approaches depends on the objectives for restoration and the circumstances in the area where restoration will occur. An important consideration is the quality of the data available for the area, particularly detailed vegetation maps and knowledge of the status and habitat requirements of species that occur there. We summarize five stages that form a logical sequence in restoration programmes and highlight some of the issues at each stage. As the outcomes of the present continent-wide experiment in restoration cannot be fully evaluated for many decades, it is prudent that a range of alternatives are trialed and monitored for their effectiveness and success.