Reduced Emissions from Deforestation and Degradation (REDD) framework demands measuring of carbon stock changes. In most cases, estimates of carbon stocks rely on volume and biomass estimation allometric models. Although species-specific and some generic models for biomass and volume estimation have been developed for some vegetation types such as Miombo woodlands, their use in Acacia-Commiphora woodlands in dry areas is questionable as they comprise short-height trees with small to medium tree diameters scattered from each other. This necessitates the need to develop robust generic allometric models that accounts for the heterogeneity of tree diversity in the Acacia- Commiphora woodlands in dry areas. Allometric models for volume and biomass estimation were developed by means of destructive sampling of 60 trees with DBH distribution ranging from 2.50 to 30.30 cm. Four different model forms were tested and the best model for estimating tree’s section biomass and volume were selected based on the lowest values of Residual Standard Error (RSE) and Akaike Information Criterion (AIC). All selected models had their parameter estimates significantly different from zero (P<0.01). The best biomass and volume models were used to estimate total tree biomass (above- and belowground) and tree volume at plot level. The estimated total tree biomass and volume was 34.69 ± 2.42 t C/ha and 23.11 ± 1.64 m 3 ha -1 respectively which is equivalent to carbon stock of 17.00 t C/ha. These estimates may be used to set a baseline for computation of carbon stock changes which are useful not only for sustainable management of the woodland but also for the implementation of REDD+ policy.