This study evaluated the capabilities of a handheld mid-infrared (MIR) spectrometer combined with multivariate analysis to characterize oils, monitor chemical processes occurring during oxidation, and to determine fatty acid composition. Vegetable oils (corn, peanut, sunflower, safflower, cottonseed, and canola) were stored at 65 °C for 30 days to accelerate oxidation reactions. Aliquots were drawn at 5 day intervals and analyzed by benchtop and portable handheld mid-infrared devices (4,000-700 cm -1) and reference methods (IUPAC 2301 [1], 2302 [1]; AOCS Cd 8-58 [2]; and Shipe 1979 [3]). PLSR and soft independent modeling of class analogy (SIMCA) models were developed for oil classification and estimation of oil stability parameters. Models developed from MIR spectra obtained with a benchtop spectrometer equipped with a 3-bounce ATR device resulted in superior discriminative performances for classifying oils as compared to those obtained from handheld spectra (single-bounce ATR). Models developed from reference tests and handheld spectra showed prediction errors (SECV) of 1 meq/kg for peroxide value, 0.09% for acid value and 2% for determination of unsaturated fatty acids in different oils. Spectral regions ~3,012-2,850 cm -1 (C-H stretching bands/shoulders of fatty acids), ~1,740 cm -1 (C=O stretching of esters), and ~1,114 cm -1 (-C-O stretching) were found to be important for prediction. Handheld-FTIR instruments combined with multivariate-analysis showed promise for determination of oil quality parameters. Portability and ease-of-use makes the handheld device a great alternative to traditional methods. © 2011 AOCS.