This study was carried out to investigate the ability of hyperspectral imaging technique in the NIR spectral region of 900–1700 nm for the prediction of water and protein contents in Spanish cooked hams. Multivariate analyses using partial least-squares regression (PLSR) and partial least squares-discriminant analysis (PLS-DA) were applied to the spectral data extracted from the images to develop statistical models for predicting chemical attributes and classify the different qualities. Feature-related wavelengths were identified for protein (930, 971, 1051, 1137, 1165, 1212, 1295, 1400, 1645 and 1682 nm) and water (930, 971, 1084, 1212, 1645 and 1682 nm) and used for regression models with fewer predictors. The PLS-DA model using optimal wavelengths (966, 1061, 1148, 1256, 1373 and 1628 nm) successfully classified the examined hams in different quality categories. The results revealed the potentiality of NIR hyperspectral imaging technique as an objective and non-destructive method for the authentication and classification of cooked hams.