The future of coral reefs is under threat since anomalous heat waves are causing the death of reef building corals around the world. This phenomenon, known as "coral bleaching", has already resulted in widespread loss of coral over the past 30 years. Without corals, the entire reef ecosystem is expected to collapse, threatening the survival of up to one third of marine wildlife. Despite the catastrophic perspectives, a glimmer of hope is brought by corals that persist at reefs exposed to recurrent heat waves. Evolutionary adaptation might underpin these observations. Characterizing the adaptive potential of corals is therefore essential, as conservation efforts could be oriented towards coral breeds resistant to future thermal conditions. To date, however, the characterization of heat adaptation in corals shows substantial gaps at different scales. At the molecular level, more needs to be discovered about the cellular pathways implicated. At a population level, there are no well-established frameworks to detect heat-tolerant corals for use in defining conservation priorities. In this research, we employed a seascape genomics approach to contribute to filling these gaps. This approach combines the environmental characterization of the seascape with a genomic analysis of populations, with the goal to identify genetic variants underpinning an adaptive role. Seascape genomics therefore constitutes an exploratory approach that portrays the adaptive potential across a population, and this facilitates the transposition of results to a conservation perspective. Furthermore, the use of genomic analyses also provides the bases to formulate hypotheses from the molecular side of adaptation. The first two articles of this thesis prepare the ground for the application of seascape genomics to corals. The first focuses on the optimization of sampling strategies. By using computer simulations, we defined the practical guidelines to organize a sampling strategy securing sufficient statistical power. The second is the proof-of-concept for the application of seascape genomics to corals. We retrieved a pre-existing genomic dataset on corals from Japan, and used it to characterize adaptive potential against heat stress and to compute conservation priorities accordingly. The third and the fourth articles of this thesis are based on new data collected in the frame of a project dedicated to the study of adaptation in three coral species from New Caledonia (Southwestern Pacific). We organized the sampling strategy using the guidelines of the first article, and processed data using the framework described in the second. The result is a dataset tailored to the needs of seascape genomics, and this enabled to refine our methods. In the third article we focus on the molecular side of adaptation and disclose cellular pathways potentially involved in heat stress adaptation. The fourth article emphasizes the role of adaptive potential under a conservation perspective. We employed pre-existing field survey data to show that reefs predicted with higher adaptive potential suffered reduced coral loss after heat stress. Together, this thesis highlights the value of the seascape genomics approach for characterizing the adaptive potential of corals and for supporting reef conservation strategies. More broadly, this work shows that building bridges from academic research to tangible conservation actions is not only possible but also essential to face this crisis.