QSAR studies have been performed on some PPAR-gamma agonists using TATA-BioSuite software to identify the essential structural and physico-chemical features for their PPAR-gamma agonistic activity. The 23 compounds were divided into training set of 18 and test set of five compounds using k-nearest neighbor (kNN) clustering. The steric, electronic and topological descriptors were found to have an important role in governing the variation in agonistic activity. The predicted activities by the developed models were in good accordance with the observed activities.