In order for accurated carbon stored study in forest system, especially in tropical forest, study in species level provides the key basic. This project concentrates on setting up regression models for volume (m3) above-ground carbon composition (trunk, branch and leave, in kg) basing on easily measured factors (Diameter of Breast Height (D1.3, cm) and total height (H, m) of 6 species: (1) Schima superba Gardn. et Champ
(2) Dacryodes dungii Dai et Jakovl (Coc da)
(3) Michelia braianensis Gagn.
(4) Sinosideroxylon bonii Aubr.
(5) Parashorea stellata Kurz. and (6) Endospermum chinense Benth. They are the main harvested species in Gia Lai evergreen broad-leaved forest, the biggiest forest area in Vietnam. Sample are 30 tree/species with diameter class is 3 cm. The results are regression equations of species volume, tree carbon (Ctr) and their parts: trunk (I), branch (B) and leave (L) with D1.3 and with D1.3 anq H. The model chosen are Power (Y = b0 x X, 1) with lowest regression coefficients, smallest error (t=0.05) and parrameters (b0, b1, t =0.05). The result also are the percentage of above-ground composition of tree (trunk, branches and leaves) of species. This is the first study on carbon distribution of individual species with easily-measuerd factors (D 1.3 and H), providing the precise basic for forest stand carbon calculation.