Quantitative Structure-Retention Relationship (QSRR) studies were performed for predicting the retention times of 43 constituents of saffron aroma, which were analyzed by solid-phase micro-extraction gas chromatography-mass spectrometry (SPME-GC-MS). The chemical descriptors were calculated from the molecular structures of the constituents of saffron aroma alone, and the linear and non-linear QSRR models were constructed using the Best Multi-Linear Regression (BMLR) and Projection Pursuit Regression (PPR) methods. The predicted results of the two approaches were in agreement with the experimental data. The coefficients of determination (R(2)) of the linear BMLR model were 0.9434 and 0.8725 for the training and test sets, respectively. The other non-linear PPR model gave a more accurate prediction with R(2) values of 0.9806 (training set) and 0.9456 (test set). The proposed models could also identify and provide some insights into structural features that may play a role on the retention behaviors of the constituents of saffron aroma in the SPME-GC-MS system. This study affords a simple but efficient approach for studying the retention behaviors of other similar plants and herbs.