<p>Mangrove communities provide vital and unique goods and services to humanity and coastal and marine systems, but they are some of the most threatened plant communities on Earth. Thus, understanding their spatial structure and distribution of plant diversity is central to devising effective conservation and management interventions. Nevertheless, studies mapping spatial structure and plant diversity in these communities remain scarce, as the existing ones have been limited to horizontal structure and field-based diversity. This study mapped plant taxonomic diversity and spatial structure of mangrove communities in a protected area in eastern Africa by employing Sentinel-2B satellite imagery. Supervised classification techniques were used to map mangrove patches using the random forest algorithm. The <i>BiodivMapR</i> R package was used to map the plant diversity. Landscape metrics were employed to map the spatial structure, specifically the&#xa0;configuration, distribution, and structural connectivity of mangrove patches. A total of 1773 mangrove patches were recorded, with a mean fractal dimension index and aggregation index of 1.75 and 96.80%, respectively. Patches with high connectivity importance were concentrated in the central parts of the study area. The estimated diversity (Shannon–Wiener Index) ranged from 0 to 1.91, with a mean of 0.54. There was a significant positive correlation between the diversity estimated by satellite images and observed in the field, with <i>R</i>² = 0.77 and RMSE = 0.0729. Although the mangroves are still in good conservation status, improving conservation efforts is needed to ease the edge effect and increase patch quality and connectivity. This study advances meaningful baselines for understanding spatial arrangement, large-scale plant biodiversity, and the quality of mangrove patches, which is crucial for developing effective conservation measures.</p>