The invention provides a method for measuring the content of an SBS modifier in SBS modified asphalt. The method comprises the following steps: preprocessing infrared spectral data of modified asphaltstandard samples with known gradient SBS contents, then substituting a training set into an artificial neural network program for training, establishing a regression model of infrared spectra and SBScontents of the modified asphalt standard samples, and then substituting similarly data-preprocessed spectral data of a modified asphalt sample to be tested, as a testing set, into the regression model to obtain a predicted value of the SBS content of the modified asphalt sample to be tested, wherein infrared spectral data preprocessing comprises wavelength selection, background subtraction, baseline correction, abnormal data rejection, unit variance processing and the like, and a nonlinear regression problem can be effectively solved through full spectral information. By the method, an erroris small, the accuracy is high, the reproducibility is good, the stability is high, the required sample amount is small, the preprocessing method is simple, and the SBS content of the modified asphalt can be quickly and effectively measured.