The concentrations of tartrazine (TAR) and allura red (ALL) in binary mixtures and commercial preparation were determined by the combined use of chemometric (or multivariate) calibrations and HPLC method. In this study, partial least squares (PLS), principal component regression (PCR) and classical least squares (CLS) based on multiwavelength HPLC data were refined as multivariate calibration techniques such as CPLS, CPCR, CCLS. The relation between multiwavelength peak area data (x-block) and concentration set (y-block) were used to obtain the chromatographic multivariate calibrations. Multiwavelength-chromatograms or multiwavelength peak area data were obtained by using photodiode array (PDA) detectors. Waters Symmetry (R) C18 Column 5 mu m 4.6 x 250 mm and a combination of 0.2 M acetate buffer (pH = 5), acetonitrile, methanol and bidistilled water (55:20:10 v/v) at the flow rate of 1.9 mL/min were used to obtain a good chromatographic separation between TAR and ALL in presence of sunset yellow (Internal Standard (IS)). These chromatographic multivariate techniques were validated by analyzing the different synthetic mixtures and by using standard addition technique. These methods were applied to the commercial soft drinking powder samples containing TAR and ALL. The results from these chromatographic multivariate techniques were compared with each other as well as obtained by alternative single HPLC method.