It is difficult to accurately determine the spatial distribution of tidal flats in intertidal zones due to periodic tidal inundation. Therefore, it is urgent to use remotely-sensed technology to detect the spectral variation characteristics of tidal flats, construct a tidal flat extraction index, and then provide methods and basic data support for flat tidal interpretation. Based on multi-temporal Sentinel-2 images, this research analyzed the spectral reflectance differences of different land cover types in the high- and low-tide images and then determined the bands that can reflect flat tidal characteristics. Finally, a tidal flat recognition index was proposed by mathematical combination. The proposed tidal flat index is studied: (1) the proposed tidal flat recognition index was applied to three study areas with different tidal flat types, and the tidal flat recognition index’s separability and applicability to different tidal flat types are studied. The results showed that the proposed tidal flat recognition index showed a good performance on tidal flat separability compared with other land cover types and is applicable to different types of sandy and muddy tidal flats; (2) the applicability of the tidal flat recognition index to different classification methods (including minimum distance method, maximum likelihood method and support vector machine) is studied. The results showed that the overall accuracy is greater than 93%, and the kappa coefficient is greater than 0.85 for distinguishing tidal flats. The tidal flat recognition index is universal to different classification methods and can effectively improve the accuracy of distinguishing tidal flats; (3) the suitability of the tidal flat recognition index to different remotely-sensed data sources is studied. Compared the Sentinel-2 images with OHS images, the results showed that the tidal flats are distinguished, and the tidal flat recognition index proposed is applicable to different data sources, achieving a higher classification accuracy. This study improves the accuracy of distinguishing tidal flats using remote sensing data, enriches the theory of flat tidal interpretation, and provides theoretical guidance and significance for the scientific management and protection of the coastal.