Global ocean warming combined with a long-lasting El Niño Southern Oscillation event (2015–-2016) are provoking the third ever global coral bleaching event. Hotspots of marine biodiversity, coral reefs require special attention from stakeholders and scientists tasked with Earth’s health. Earth observation has recently benefited from the launch of the multispectral Sentinel-2 (S2) sensor delivering high spatial resolution (10 m) optical information worldwide every 10 days. Within the Sen2Coral project of ESA, and as a precursor of open-source specific data processors’ development, this study strives to evaluate in real conditions the S2 capabilities for coral reef bathymetry and habitat mapping as well as bleaching detection given the ongoing event. A dedicated campaign held in Fatu Huku (Marquesas Archipelago, French Polynesia) in February 2016 enabled geolocated water depths and benthic frames to be extensively collected. After geometric and radiometric standardization, two sets of S2 blue, green, red and near-infrared bands were processed for comparison (11 February and 11 April 2016). The bathymetry mapping up to 20 m was satisfactorily generated for both dates (R2=0.64, r=0.79, RMSE=0.12 and R2=0.74, r=0.86, RMSE=0.09, respectively). The benthic composition (deep water, sand, reef pavement and live reef) was also reasonably classified (OA=78.8%, κ=0.71 and OA=73.5%, κ=0.65). Bathymetric maps and difference maps of benthic composition and depth-invariant reflectance were used to evaluate changes in water surface, water column and benthic habitats of this oceanic lagoon. By focusing on areas deprived of water-related constraints, benthic shifts from live reef to reef pavement were locally pointed out suggesting a potential bleaching detection at this 10-m scale when fit in the radiometric error budget. Those findings hold great promise to use S2 for coral reef bathymetry and habitat mapping but also are an incentive to confirm bleaching detection over larger areas provided with denser ground-truth data, unequivocally correlating the bleaching event over S2 data.