Teak plantations are widely promoted as productive forest systems with potential contributions to carbon storage and climate-oriented land management. However, plantation carbon estimates often rely on generic biomass-to-carbon conversion factors, which may overlook variation in carbon concentration among tree components. This study refined carbon estimates in tropical teak plantations in Nayarit, western Mexico, by combining pantropical allometry with measured carbon fractions. Carbon concentration was determined in leaves, branches, stem, roots, and total biomass, and carbon stocks were compared using the generic 0.47 factor and measured total biomass carbon fractions. Carbon concentration differed among biomass components, with leaves and branches remaining below 47%, while stem, roots, and total biomass exceeded this value. The measured total biomass carbon fraction averaged 48.24%, producing refined carbon estimates that were consistently higher than those obtained with the generic factor. Across plantation-age records, the refined approach increased carbon estimates by 1.33 Mg C ha⁻¹, equivalent to a mean relative adjustment of 2.67%. When projected as an illustrative scenario, this difference represented 133, 665, and 1330 Mg C over 100, 500, and 1000 ha, respectively. These findings show that measured carbon fractions can reduce one source of conversion-related uncertainty and refine plantation-level carbon estimates. Broader regional application would require larger and more representative plantation inventories.