Seagrass are marine flowering plants that form extensive meadows in shallow coastal waters. They play a critical role in coastal ecosystems by providing food and shelter for animals, recycling nutrients, and stabilizing sediments. Therefore, they are widely used as an ideal biological indicator for assessing the health status and quality of coastal ecosystems. In the Alfacs Bay (Ebro Delta), seagrasses are located in the shores, showing an annual variation with a peak in summer. The decreasing of averaged salinity and increasing of nutrients concentration and turbidity, has led to a notable reduction of the seagrass beds. Thus, a cartography to monitor spatiotemporal changes of meadows and to forecast the evolution of the environmental characteristics of the system, is needed. Nowadays, the standard methodology is a combination of photointerpretation and field prospection with significant workload resources. In contrast, an automatic methodology relying on multispectral moderate resolution Sentinel 2 (S2) satellite imagery is proposed. The methodology consists of: atmospheric correction of Level-1C images, application of Green Normalized Difference Vegetation Index, statistic thresholding to tell apart possible seagrass areas and a supervised learning method to refine this classification and to identify habitats. The methodology has been applied and calibrated using S2 satellite imagery and reference data comprising several patches distributed along the Alfacs Bay. In these patches, seagrass areas were identified (visually and location with GNSS). The results showed that seagrass meadows can be automatically delineated using S2 imagery.