Mangrove forests are highly productive ecosystems that play a central role in coastal carbon dynamics, yet aboveground biomass across much of the Amazon coast remains poorly quantified. To address this gap, species-specific allometric equations were developed for Rhizophora mangle, Avicennia germinans, and Laguncularia racemosa on the Ajuruteua Peninsula in the northeastern Amazon. Forest structure was characterized across multiple plots, and destructive sampling supported the formulation of equations for total and compartmental biomass. Diameter, height, and wood density were evaluated as predictors, with diameter at breast height as the most informative variable. The selected logarithmic models showed strong performance and robustness, enabling reliable biomass estimation across species and size classes. Biomass partitioning patterns underscored the dominance of woody compartments and the influence of environmental gradients on forest structure. These models strengthen carbon stock assessments and provide a critical methodological basis for blue carbon strategies in Amazonian mangroves.