1. Restoration ecology is moving towards designing restoration actions to maximize ecosystem services (ES). Such restoration actions require planning at large spatial scales, as these are often more meaningful for ecosystem functioning and ES supply. As economic resources to undertake ecological restoration at large scales are scarce, prioritizing sites to enhance multiple ES supply is critical. 2. Our study presents an index, the Relative Aggregated Value of ES (RAVES), to prioritize sites for ecological restoration based on the assessment of multiple ES. We tested the spatial heterogeneity of ES to identify the relevant scale to managing ES and to applying the RAVES index using a local case study. We also used the RAVES index to compare three alternative restoration scenarios to enhance ES based on the availability of socio-economic resources. 3. The highest RAVES values were found in areas with natural vegetation and in gorges with riparian forests. The lowest RAVES values were found in crop fields, steep slopes, and river stretches without riparian forest. 4. The multi-scale spatial analysis indicated that most ES showed significant heterogeneity at multiple spatial scales, especially at broad (20–30 km) and very broad (40–50 km) scales. However, at spatial scales smaller than 2 km, only biological control showed significant heterogeneity. 5. The optimal socio-economic conditions to enhance ES supply were met in a scenario where both private and public land and economic funds were available to implement ecological restoration. As most areas with low RAVES were in private lands, even with scarce economic funds restoration of private lands would result in a large increase of RAVES. 6. Synthesis and applications. The RAVES index is a practical tool to hierarchically prioritize sites for ecological restoration across large spatial scales. When linked to scenario analysis, the RAVES index can also be used to identify optimal management scenarios. By linking ES assessments to the identification of the spatial scales at which ES show largest heterogeneity, our analyses can help decision-makers identifying the spatial scale at which ES are more likely to respond to management and highlight the need to integrate local and regional management plans.09-Nov-2017