The potential for large-scale afforestation, as a climate change mitigation strategy, has led to national afforestation targets being proposed. These targets raise many questions, including how to optimise carbon storage, whilst enhancing other ecosystem services (ES). Here, we assess how decision-making at different scales might affect the spatial distribution of new tree planting in England and hence ES provision. This was achieved by modelling the impact of afforestation on three example ES (carbon sequestration, recreation, and flood mitigation) to identify the best areas for planting at five scales of decision-making: national, regional, county, district, and parish. The modelling shows that carbon sequestration rates are relatively invariant across England (once unsuitable areas are excluded), whilst recreation and flood mitigation are more spatially variable, so full simultaneous optimisation of the three ES is not possible. Consequently, recreation and flood mitigation are also more impacted by changes in planning-scale than carbon, with the modelling showing over 200% difference between the parish and national-scenarios for flood mitigation benefits. Overall, targeted afforestation at national scale maximises ES benefits, but risks overwhelming some landscapes with trees. Targeting afforestation using smaller planning units generates more evenly distributed planting, but lower ES benefits. Our results show national-scale planning produces over 35% more combined ecosystem value than planning at parish-scale, and over 65% more than randomized planting. This highlights the benefits of targeting at a relevant scale, but also the potential trade-offs between widely-distributed new tree planting (e.g., parish-scenario), versus planting to optimise ES provision (e.g., the national-scenario). These results do not imply a single ‘correct’ scale for planning, but instead highlight the importance of considering the impact of a specific scale on the different ES, and that even spatial targeting at the smallest scale produces higher ES benefits than random planting.