The gray mangrove [Avicennia marina (Forsk.) Vierh.] is the most widely distributed mangrove species, ranging throughout the Indo-West Pacific. It presents remarkable levels of geographic variation both in phenotypic traits and habitat, often occupying extreme environments at the edges of its distribution. However, subspecific evolutionary relationships and adaptive mechanisms remain understudied, especially across populations of the West Indian Ocean. High-quality genomic resources accounting for such variability are sparse. We sequenced and assembled the genome of A. marina from the Arabian Gulf, which is the harshest region that the species occupies and at the northern-most limit of its distribution. We used proximity ligation libraries Chicago and Dovetail HiC, and the HiRise assembly pipeline, producing a 456,556,596 bp long genome. The largest 32 scaffolds (22.4 Mb to 10.5 Mb) accounted for 98 % of the genome assembly, with the remaining 2% distributed among much shorter 3,777 scaffolds (62.4 Kb to 1 Kb). We annotated 23,331 protein-coding genes using tissue-specific RNA-seq data, from which 13,312 were associated to GO terms. Genome assembly and annotated set of genes yield a 96.7% and 92.3% completeness score, respectively, when compared with the eudicots BUSCO dataset. Furthermore, an FST survey based on resequencing data successfully identified a set of candidate genes potentially involved in local adaptation, and revealed patterns of adaptive variability correlating with a temperature gradient in Arabian mangrove populations. Our A. marina genomic assembly provides a highly valuable resource for genome evolution analysis, as well as for identifying functional genes involved in adaptive processes and speciation. ### Competing Interest Statement The authors have declared no competing interest.