Plants in arid areas are widely used for timber, food, fire wood and livestock fodder. As a result they could be threatened by over-exploitation from overgrazing and overharvesting. These threats are exacerbated by lack of knowledge on the composition and population structure of woodland species. The aim of this study was to classify woodland species and determine their population structure. Sampling was conducted in 20 × 50m plots. In total 75 plots were sampled where canopy cover was estimated and diameter at breast height measured. Hierarchical cluster analysis was used to derive different communities. Indicator species analysis was utilized to determine the indicator value for different species in a given community. Plant species richness and diversity were determined for each plant community. Multi Response Permutation Procedures (MRPP) was used to determine the level of separation between the plant communities. Kruskal-Wallis test was exercised to compare species diversity and richness between woodland communities. Plant species were classified into four communities of Acacia erioloba-Grewia bicolor, Terminalia sericea-Burkea africana, Acacia mellifera-Boscia albitrunca and Combretum collinum-Baphia massaiensis. Results of MRPP showed that the communities were significantly (p<0.05) different from each other. Species richness and diversity were significantly lower in Acacia mellifera-Boscia albitrunca community. Populations of Acacia erioloba, Philenoptera nelsii and A. nigrescens showed bell-shaped structure while those of Burkea africana, Boscia albitrunca and Combretum collinum had reverse J-shaped pattern. Terminalia sericea, Ziziphus mucronata and Colophospermum mopane had only 2 or 3 classes skewed towards juveniles.