A novel algorithm for component identification in mixtures based on Raman spectroscopy is proposed, in which background correction and de-noising operation on the Raman spectra of mixtures are firstly performed, the Voigt function is then used to fit the Raman peaks to obtain the Raman shifts, full widths at half maximum, and peak intensities as the feature parameter vector of mixtures, and finally, the effective component identification of mixtures is finally realized via the correlation analysis of the feature vectors between the mixtures and the pure substances in the database. The Raman spectral data of 18 pure substances are used to build a standard database, and the component identification experiment of 6 kinds of mixtures are conducted. The experimental results show that the recognition accuracy of the proposed algorithm is up to 100%.