The invention discloses a seawater pollution area identification method and device based on a high-resolution remote sensing image, and belongs to the field of digital image processing. The method comprises: firstly, performing sea-land automatic classification on remote sensing images through a supervised learning algorithm, a classification result can reach a high precision level through processiterative clustering, and meanwhile, compared with an existing sea-land boundary analysis and classification method, the operand being small; then, using the chlorophyll concentration difference andthe brightness difference of pollutant shadows existing between the marine pollution area and surrounding seawater; combining a normalized vegetation index related to chlorophyll, a normalized water body shadow index related to brightness, a segmentation-based image interpretation thought and a human eye vision-based saliency mechanism in remote sensing interpretation; through threshold segmentation, realizing extraction of a marine pollution area, respectively extracting an area with excellent water quality and a severely polluted area, and further obtaining a pollution transition area. The method provided by the invention provides convenient and accurate reference for marine pollution prevention and treatment.