The feasibility to extract the polyphenols spectra information from near infrared spectroscopy of green tea infusion using chemometrics methods was attempted. A cylindrical quartz glass tube with an optical path of 5mm was used. First, synergy interval PLS (siPLS) was implemented to select efficient spectral regions from SNV preprocessed spectra. Then, optimal variables were selected using genetic algorithm (GA) from these selected spectral regions by siPLS. This study showed that the selected characteristic spectral intervals were not within the range of the strong absorbance for water. When 166 variables selected by GA and eight PLS factors were included, the optimal model (siPLS-GA) was achieved with PRMSEP=0.685%, PRSD=5.26% and PRPD=3.22 in the prediction set. The results showed that the performance of siPLS-GA model could be employed to measure polyphenols content in green tea infusion by near infrared spectroscopy in practice.