随着遥感数据量的爆发式增长，对变化过程分析的精细化要求与本地算力不足之间的矛盾日益突出。GEE（Google Earth Engine）地理云平台的出现，解决了用户算力紧张的行业痛点。本文以粤港澳大湾区为研究区，在GEE的支持下，构建1987年—2020年年度湿地分类数据集，分析大湾区红树林的时间阶段性特征和空间扩张过程，结合连续长历时分析对变化时间点的准确识别，揭示了保护区设立与滩涂造林等工程在红树林保护与修复中的积极成效。主要结论有：（1）截止到2020年，大湾区共有红树林2174 ha，81%的红树林集中在深圳湾、淇澳岛和镇海湾；（2）大湾区的红树林经历了由平稳发展（1987年—2003年）到快速增长（2003年—2020年）的变化过程，其主要增量来源于镇海湾（40%）和淇澳岛（28%）；（3）淇澳岛和镇海湾的红树林仍处于快速增长期，但淇澳岛增速最快，从2002年至今面积翻了30倍，深圳湾则在早期的快速增长（1987年—2009年）后进入平稳期（2009年—2020年）；（4）由于保护区设立时间较早，深圳湾成为大湾区唯一一个形成稳定核心区的红树林分布区，镇海湾虽然拥有最大的红树林面积，但林道狭窄，景观破碎，生态系统反而更加脆弱；（5）设立自然保护区和滩涂造林都对红树林面积增长起到了积极作用。本研究为大湾区海岸带湿地生态系统保护与修复提供科学的证据支持，对沿海生态屏障建设具有一定的指导作用。

With the explosive growth of remote sensing images, the contradiction between the refined requirements of the change processing analysis and the lack of local computing power has become increasingly prominent. The emergence of Google Earth Engine (GEE) geographic cloud platform has solved the pain points of the industry, where users are strained with computing power.The Guangdong-Hong Kong-Macao Greater Bay Area (GHM) is considered the study area, and an annual wetland classification data set from 1987 to 2020 is constructed with the support of GEE. It is analyzed for the temporal phase characteristics and spatial expansion process of mangroves, and the result reveals the positive effects of the establishment of protected areas and tidal flat afforestation on the mangrove protection and restoration by combining the accurate time point of change identified by the continuous long time series analysis.The imaging quality of optical images is greatly reduced because the study area is located in a tropical and subtropical cloudy and rainy area. To fully use the satellite remote sensing multitemporal features to effectively eliminate clouds/shadows, all Landsat images during the time period are collected to form a continuous time series. With the support of the GEE cloud platform to satisfy the requirements of computing power for the large amounts of data, the random forest algorithm is applied to obtain the annual wetland classification data set from 1987 to 2020. On this basis, the spatio-temporal characteristics of mangrove over long time are mined.According to the accuracy assessment, the averaged out-of-bag error of wetland data set is 6.61%±0.08%, and the average identification accuracy for mangroves is 89.64%±0.13%. In 2020, the Greater Bay Area has 2,174 ha of mangrove forests, and 81% of the mangrove forests is concentrated in Shenzhen Bay, Qi’ao Island, and Zhenhai Bay. The mangrove forests in the GHM experience steady development (1987—2003) and then rapid growth (2003—2020), the main increase is observed in Zhenhai Bay (40%) and Qi’ao Island (28%). The mangrove forests in Qi’ao Island and Zhenhai Bay are still in a period of rapid growth. However, Qi’ao Island has the fastest growth rate; it has doubled its area by 30 times since 2002. Shenzhen Bay has entered a stable period (2009—2020) after its early rapid growth (1987—2009). Shenzhen Bay became the only mangrove distribution area that formed a stable core area in the GHM due to the early establishment of the reserve. Although Zhenhai Bay has the largest area of mangrove forests, the ecosystem is more fragile because of the narrow width and the fragmented landscape.The establishment of nature reserves and tidal flat afforestation has played an important role in the growth of mangrove area. Integrated monitoring, such as satellites, drones, and ground monitoring, should be incorporated into the mangrove protection and restoration assessment. This study provides scientific evidence support for the implementation of the sustainable development strategy goals of the GHM and has a certain guiding role in the construction of coastal ecological barriers.