Biological interactions that involve the capture of the ever increasing levels of carbon dioxide in forest ecosystems have drawn global attentions to carbon accounting. In the present study the desirable predictor variable in carbon accounting was established in Pseudolachnostylis maprouneifolia Pax. Twenty one trees of P. maprouneifolia Pax with diameter at breast height (DBH) from 5.7 to 14.7 cm and total height (H) between 300 cm to 672 cm were destructively sampled from Miombo woodland of Ilunde in Tanzania. Regression analysis was carried out by using DBH and H as predictor variables. The best fit allometric models in stems and branches were obtained when both H and DBH were used as predictor variables with coefficient of determination values (R2) of 0.99 and 0.79 at P < 0.05, on the contrary, in leaves and/or twigs the best fit allometric model had R2 of 0.96 when DBH was the only predictor variable.In P. maprouneifolia Pax. measurement of both DBH and H is recommended for reliable carbon inventory in the Central Zambezian Miombo woodlands under the specified tree dimensions.World Bank through Mkwawa University College of Education