Quantitative structure activity relationships (QSAR) were developed relating the 14th and 21st day antituberculosis activity against H37Rv strain of Mycobacterium tuberculi and antibacterial activity against Staphylococcus aureus ATCC3750, Bacillus subtilis 6633, Escherichia coli ATCC3750, and Salmonella typhi NCTC786 of 55 pyrazine containing thiazolines and thiazolidinones, with the molecular descriptors. The developed models were able to fit the data well (r2 = 0.69-0.87) and had reasonable predictive capability (q2 > 0.62). The data were also divided into a training set and a test set, the former was used to develop the QSAR and the latter was used to evaluate the predictive capability of these developed models. In all the cases, the models were able to predict the test data set reasonably well. Predominantly, pyrazine ring is well-known for its antimycobacterial activity and hence these equation could be used to design newer analogues with higher activity. These compounds also possess both antitubercular and antibacterial activity. Descriptors pertaining to electronic, topology, and hydrophobicity of the molecules appear in the model equations. © 2008 The Authors.