The objective of this study was to evaluate the performance of a portable FT-IR spectrometer equipped with a 5-bounce heated ZnSe crystal to develop classification methods for the authentication of potato chip frying oils and to generate prediction models for monitoring oil quality parameters for real-time and field-based applications. Oil from commercial potato chips (n = 95) was expelled mechanically by a hydraulic press and the fatty acid profile determined by GC-FAME to identify the oil type used for chip manufacturing. The peroxide value (PV), free fatty acids (FFAs), and p-anisidine value (p-AV) were also evaluated to determine quality parameters of the oils. IR spectra were collected using a portable FT-IR equipped with a heating stage (65°C) and analyzed by pattern recognition using a soft independent modeling of class analogy algorithm (SIMCA) and partial least squares regression (PLSR). SIMCA showed that different oil types successfully formed distinct clusters allowing detection of the mislabeling of frying oils in commercial chips. PLSR models predicted the fatty acid profile (GC-FAME) with excellent correlation (Rcal ≥ 0.93) and the standard error of cross-validation (SECV) of ∼1.0% for major fatty acids. The models for FFAs, PV and p-AV gave an Rcal ≥ 0.93 and SECV of 0.05%, 1.27 meq kg-1, and 5.94 p-AV, respectively. Profits and trading advantages from mislabeling prejudice consumers and manufacturers, and our data supports that IR portable instruments present great potential for in situ surveillance of vegetable oils used for potato chip frying.