Laser-induced breakdown spectroscopy (LIBS) usually contains rich element characteristic spectrum peaks. The accurate identification of spectral peaks is the premise and basis for qualitative and quantitative elements analysis in LIBS technique. In this paper, a new method for automatic identification of spectral peaks is proposed to overcome the disadvantages of low accuracy and reliability of traditional spectral peak recognition methods. Firstly, the Voigt function is used to fit the experimental spectrum to overcome the overlapping spectrum and noise interference. Then, the center wavelength, light intensity, full width at half maximum and peak centroid of the fitting spectrum are extracted as the characteristic parameter vector of spectral peak after fitting process. Finally, the similarity analysis of the characteristic parameter vector of spectral peak between the spectrum to be recognized and the spectrum in National Institute of Standards and Technology (NIST) standard spectral database is carried out, so as to realize the automatic recognition of the spectral peak corresponding elements. Experiments are carried out using NIST standard database and LIBS spectra of tea samples, and the effectiveness of the method for spectral element recognition based on LIBS is verified.