In this thesis study, surface cleaner samples were studied. UV-Visible Spectrophotometry was applied to the samples prepared by adding dye amounts determined at various concentrations into these samples. Partial least squares (PLS), simple least squares regression (SLR) and genetic inverse least squares (GILS) methods have been successfully applied to the visible spectra of prepared samples for the quantitative determination of these coloring substances. A total of 35 samples were prepared as binary and ternary mixtures of these dyes along with single dye containing cleaning products. Among these 35 samples, 29 of them were chosen as calibration set and the remaining 6 samples were used as independent validation set. Absorption spectra between 400-700 nm were recorded on the prepared set and SLR, PLS and GILS methods were applied to the obtained spectral data and the resulting predictions were compared with actual values. The regression coefficients (R²) and standard error cross-validation (SECV) values of these methods (PLS and GILS) and the standard error of prediction (SEP) were found, the results were evaluated, the accuracy of the chemometrics modelling was reviewed, and the predictions and references were compared. With these analysis methods, the concentration of dyes in the cleaning products can be determined by using chemometrics methods combined with UV-Visible Spectroscopy.