We developed species specific equations to predict aboveground biomass (AGB) of ten woody species in Borana rangelands of southern Ethiopia. A total of 150 plants 15 for each species were measured for biometric variables including the diameter at stump height (DSH), diameter at breast height (DBH), tree height (TH) and crown diameters were destructively harvested to obtain dry biomass. Many equations that related three biomass components: total aboveground, stem and branches to single or combination of predicator variables: DSH, DBH, TH, crown area (CA) and crown volume (CV) fit the data well to predict total AGB and by components for each of the species (adj.R2 > 0.80; P 0.80; P 0.93; P < 0.0001. A generalized mixed-species allometric model developed from the pooled data of seven species was most accurately predicted by the combination of three predicators (DSH-TH-CA models), with adj. R2 between 0.84 and 0.90 for all AGB categories. Hence, our species-specific allometric models could be adopted for the indirect biomass estimation in semi-arid savanna ecosystem of southern Ethiopia. The mixed species allometric models will give a good opportunity when species-specific equations are not available and contribute to estimate the biomass and carbon stock in woody vegetations of East African rangelands.