Rehabilitation and restoration of forest ecosystems are in growing demand to tackle climate change, biodiversity loss, and desertification—major environmental problems of our time. Interest in the restoration of ecosystems is increasingly translated into a strong political commitment to large-scale tree planting projects. Along with this new impetus and the enormous scale of planned projects come both opportunities and risks: opportunities to significantly increase the use of native species, and risks of failure associated with the use of inadequate or mismatched reproductive material, which though it may provide forest cover in the short term, will not likely establish a self-sustaining ecosystem. The value of using native tree species in ecosystem restoration is receiving growing recognition both among restoration practitioners and policymakers. However, insufficient attention has been given to genetic variation within and among native tree species, their life histories, and the consequences of their interactions with each other and with their environment. Also, restoration practitioners have often neglected to build safeguards against the anticipated effects of anthropogenic climate change. Measurement of restoration success has tended to be assessments of hectares covered or seedling survival in a short timeframe, neither of which is an indicator of ecosystem establishment in the long term. In this article, we review current practices in ecosystem restoration using native tree species, with a particular focus on genetic considerations. Our discussion is organized across three themes: (i) species selection and the sourcing of forest reproductive material; (ii) increasing resilience by fostering natural selection, ecological connectivity, and species associations; and (iii) measuring the success of restoration activities. We present a number of practical recommendations for researchers, policymakers, and restoration practitioners to increase the potential for successful interventions. We recommend the development and adoption of decision-support tools for (i) collecting and propagating germplasm in a way that ensures a broad genetic base of restored tree populations, including planning the sourcing of propagation material of desired species well before the intended planting time; (ii) matching species and provenances to restoration sites based on current and future site conditions, predicted or known patterns of variation in adaptive traits and availability of seed sources; and (iii) landscape-level planning in restoration projects.