Simple, fast, and accurate analytical techniques for verifying the accuracy of label declarations of marine oil dietary supplements containing eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) are required due to the increased consumption of these products. We recently developed broad-based partial least squares regression (PLS-R) models to quantify six fatty acid (FA)/FA classes using the spectroscopic data from a portable Fourier-transform infrared (FTIR) device and a benchtop FT-near infrared (FT-NIR) spectrometer. In this study, we propose an improved quantification method for these FA/FA classes by incorporating a non-linear calibration approach based on the machine learning technique support vector machines (SVM). For the two spectroscopic methods, higher accuracy in prediction was observed as indicated by low root mean square error of prediction (RMSEP) values and correlation coefficient values (R 2 ) close to one indicating excellent model performances. The percent accuracies of the support vector regression (SV-R) model predicted values for EPA and DHA for a reference material were from 90 to 110%. In comparison to PLS-R, signiﬁcant enhancement of SV-R accuracy in prediction of FA/FA class concentrations were obtained: up to 2.4 times for both ATR-FTIR and FT-NIR spectroscopic data. Further, SV-R models yielded better agreement with the certified/reference values for the prediction of EPA and DHA in the reference standard. Based on our findings, the SV-R methods were found to be superior in terms of accuracy and predictive quality in predicting the FA concentrations of marine oil dietary supplements. The combination of SV-R with ATR-FTIR and/or FT-NIR spectroscopies can potentially be applied to the rapid screening of marine oil products to verify the accuracy of label declarations.