The aim of this study was to predict the total antioxidant capacity (TAC) of altiplano colored quinoa using visible and near-infrared (Vis-NIR) spectral data and chemometrics. In colored quinoa, at least part of the phenolic compounds are included in the non-extractable fraction and therefore the TAC should be measured by a non-conventional QUick, Easy, New, CHEap and Reproducible procedure (QUENCHER). In this work, QUENCHER using 2,2-diphenyl-1-picrylhydrazyl (DPPH) returned a value 1.03–7.03 times higher than the extractable antioxidant capacity, and it represented 2.8%–85.8% of the TAC of altiplano colored quinoa. We applied partial least squares regression and ten-fold cross-validation to the raw and pre-processed NIR spectra for the calibration and validation of models, to predict the TAC of altiplano colored quinoa. The best model was obtained for whole quinoa seeds as opposed to ground samples (calibration: R2 = 0.97, RSME = 2.9; validation: R2 = 0.73, RSME = 8.6) using Savitzky-Golay smoothing, the D-trend method and multiplicative scatter correction procedures. The QUENCHER-DPPH result was also positively correlated with extractable antioxidant capacity (EAC-DPPH), total free phenolics, total betalain content and L values (p ≤ 0.001). Therefore, we demonstrate that the TAC of colored quinoa seeds, including both non-extractable and extractable antioxidants, can be predicted by Vis-NIR spectra coupled to chemometrics.