A method of determination of multiple food colorants in carbonate beverages from ultraviolet-visible(UV-Vis)spectrum data using kernel independent component analysis(KICA)and support vector regression(SVR)was proposed.The UV source spectra profiles of food colorants dissolved in water were estimated using KICA,and the quantitative model were established by support vector regressing(SVR)using the ICs coefficients.Adding different contents level of standard colorants to the beverage,then the information of ICs and their coefficients matrix can be obtained by KICA.In this paper,the multiple food colorants in the beverage can be estimated from the UV data using the established UV-KICA-SVR model.This method can be used to simultaneously determine multiple food colorants,i.e.lemon yellow,sunset yellow,red temptation,amaranth,carmine,and brilliant blue,with the relative standard deviations(RSDs)are 1.5%,2.2%,2.0%,2.5%,2.6%,and 1.2%,respectively,and the limits of detection(LOD)are 0.5,0.5,0.5,1.0,1.0 and 0.5 mg·L-1,respectively.