reliable estimates of biomass for the vegetation of the arid chaco forest is critical for its management, conservation or restoration. however, simple and non destructive techniques are not largely available. this study provides regression models aimed at estimating the aerial biomass of native shrub and tree species from the arid chaco, belonging to the fabaceae family (prosopis flexuosa, geoffroea decorticans, cercidium praecox, acacia furcatispina, mimoziganthus carinatus and prosopis torquata). an additional objective was to explore the feasibility of using a single equation for several species. the wood density differed significantly between species. linear and nonlinear regression models of aerial biomass were adjusted for species, with different combinations of base diameter (dab) and height as independent variables. the models properly predicted the biomass by species, the most appropriate being the linear regression model using dab2 as an independent variable. to establish differences in the behavior of species and, more precisely, in the ratio between total dry weight and dab, the slope homogeneity was analyzed using a regression model with auxiliary variables. taking into account the results obtained and the basic wood density and morphological growth structure, four groups of species have been postulated, and it was feasible to apply a single equation per group. the proposed groups are: (1) monopodial hardwood (geoffroea decorticans and cercidium praecox), (2) multi-stem shrubs of extremely heavy wood (acacia furcatispina and mimoziganthus carinatus), (3) monopodial of very heavy wood (prosopis flexuosa) and (4) monopodial and multi-stem of extremely heavy wood (prosopis torquata).