Caribbean coral populations are actively being restored via asexual fragmentation of adult populations and the production of sexual offspring. The goal of restoration is to increase population sizes while maximizing adaptive potential of restored populations. Because the total number of remaining wild coral genotypes far exceeds the capacity of nurseries to propagate them all, we provide guidelines for selecting species and genotypes for restoration projects. For each species selected, we suggest that collecting 3-4 genotypes per reef is sufficient to capture much of the allelic diversity. Because environmental conditions are rapidly changing, a mixed provenance strategy is preferred where genotypes are sourced both locally as well as from more distant sites within the management area, allowing for the inclusion of potentially adaptive genetic variants under a wider range of environmental gradients. We further recommend that nurseries concentrate on propagating genotypes with records of high relative growth, outplant survival, bleaching and infectious disease tolerance or resistance, and successful sexual propagation with other genotypes (“winners”). At the same time, some low-level propagation effort for genotypes not performing well in nurseries (“runts”) should be maintained to guard against unintended selection during captivity and to preserve genetic variants that could become adaptive in the future. Given current evidence, we do not think that outbreeding depression is a major problem. Likewise, we do not expect inbreeding depression to be an issue because currently only F1 and F2 larvae are raised. We stress that high rates of sexual reproductive success will be essential for adaptation. Hence, our recommendations outline tools to optimize coral fitness while providing adequate genetic diversity for restored coral populations to rapidly adapt to changing environmental conditions.