Restoration can be an effective measure to counteract declines of reef-building coral populations. Besides decades of research and practice, very little to no emphasis has been placed on how the adequate selection of restoration sites can aid in the recuperation of ecologically critical processes such as connectivity. Combining image classification of drone-derived aerial data with geoinformatics, we aim to identify restoration sites that exhibit the highest potential to increase the structural connectivity and fertilization potential of coral keystone species, with a particular focus on Acropora palmata. We apply photogrammetry and Object-Based Image Analysis to map the substrate of two natural reefs with contrasting coverage and spatial distribution of the species, including a hypothetical reef with a random distribution as a comparative baseline to represent a highly structurally degraded system. Priority restoration sites at each reefscape are then identified through spatial modelling using three connectivity metrics from classical landscape ecology. Our modelations suggest that restoration sites connecting or bordering major patches of A. palmata have the greatest potential to increase structural connectivity. Reefs of more degraded status are favourable because they exhibit a greater increase in connectivity metrics per area restored, while also maximizing the fertilization potential between outplanted and existing colonies. Furthermore, the spatial extent that needs to be restored to achieve maximum efficiency is greatly dependent on the initial coverage and distribution of the focal species and should, as such, always be considered in initial and repetitive restoration activities. We show the importance to include spatial planning in the site selection process of coral reef restoration and provide a methodological framework that can aid in tailoring related strategies in accordance with the spatial arrangement of the target species.