Abstract: The coastline and tidal flats of bay play crucial roles in ecology and are valuable for resources. This paper used the Google Earth Engine (GEE) for batch creation of multi-feature datasets to address the intricate environmental conditions and the regular tidal variations in bays. By comprehensively utilizing K-means clustering, decision trees, and diverse post-processing techniques, a method was proposed to automatically extract annual spatial information of the coastline and tidal flats, and typical bays with notable changes were selected for application verification. With the MovMean algorithm developed in the R language, the inter-annual evolution of the coastline and tidal flat shape of Jinghai Bay was quantitatively depicted and characterized from 1985 to 2023. It showed that over the past 39 years, the length of the coastline of Jinghai Bay had increased from 93.17 km to 133.82 km. The coastline had generally become more tortuous, with most of it expanding towards the sea, especially in the northern part of the bay (129.73 m/a). Significant changes and consistent trends had been observed in the bay and tidal flats, resulting in losses of 49.32 km² (41.61%) and 11.89 km² (35.96%), respectively. Coastline shape showed a phase of small changes (1985−2001), followed by significant changes (2001−2015), and then moderation (2015−2023), which was consistent with the evolutionary characteristics of the national bays.