Due to the biological effects of phenolic acid components, polyphenol-rich foods are a significant part of human and animal diets. In this study, the acidity constants of caffeic acid (3,4-dihydroxycinnamic acid) in binary mixtures of ethanol–water were determined spectrophotometrically using the introduced net analyte signal (NAS) algorithm and an ab initio quantum mechanical method. The NAS is an efficient chemometric algorithm for analysis of acid–base equilibrium systems by a spectrophotometric method. At different pH values, the distribution of acid species is obtained from an absorption data matrix and this procedure enabled the pKa of caffeic acid to be obtained alternatively. The results showed that pKa1 (4.02, 4.26, 4.39, 4.57 and 5.11) and pKa2 (8.43, 8.68, 8.79, 9.00 and 9.34) were increased by increasing the percent ethanol in water (0, 10, 20, 30 and 40 vol. %) and these results were in agreement with the results of the Gaussian method. The ab initio calculated Gibbs energy change showed that para-hydroxy group is more acidic than meta-hydroxy group. The red shifts of different species of caffeic acid obtained using the ab initio quantum mechanical method are in good agreement with the results of UV–Vis spectroscopy.