In order to improve the detecting precision and robustness in determination of the content of free amino acid in tea by the FT-NIR spectroscopy,a predictive model was established by selecting efficient spectral regions combined with the Partial Least-squares(PLS) algorithm.In this research,the synergy interval Partial Least-squares (si-PLS) and the Genetic Algorithm Partial Least-squares (GA-PLS) were applied to select the efficient spectral regions.The number of PLS components and the number of intervals were optimized according to the cross-validation in a calibration set.The performance of the final model was evaluated according to the Root Mean Square Error of Prediction (RMSEP) and the Correlation coefficient (R) in prediction and calibration sets.Experimental results show that the two methods are able to produce better prediction models in relation to the full-spectrum model,and the performance of siPLS model is better than that of the GA-PLS.The optimal model offers its R in 0.954 2 and RMSEP in 0.256 0 respectively by a prediction set.This study demonstrates that NIR spectroscopy combined with efficient selection wavelength region algorithm can be used successfully to analyze the content of free amino acid in tea.