A QSRR study was performed to develop a predictive model correlating molecular structures with their RIs in gas chromatography. CNS agents included some benzodiazepines, barbiturates and phenytoin, which had been studied, previously, on two DB stationary phases with different phenyl percentages. At first, a high quality model (R 0.988, F 253.89 for DB-5 and R 0.983, F 164.721 for DB-17 column) was generated for two columns separately, using only calculated descriptors and multiple linear regression techniques. Then a combined model, added a polarity term of stationary phase (M), was also developed for these columns, and the result was satisfactory (R 0.989, F 232.354). The results of leave-one-out cross and external validation techniques showed good predictive ability of the models (R 2 cv 0.967, R 2 Prediction set 0.954). The results of the study indicated that a sixparameter equation can be utilized for prediction of retention indices of compounds on DB-5 and DB-17 stationary phases simultaneously, for which there are no empirical RI values.