To establish near-infrared spectrometry quantitative models for the determination of asiaticoside in Centella Total Glucosides.Methods：The contents of asiaticoside of 66 samples of Centella Total Glucosides were determinated by high-performance liquid chromatography.And the near-infrared spectrogram of 66 samples of Centella Total Glucosides was collected at the same time.After smoothing pretreatment by Multiple scattering correction(MSC)，the second derivative(2D)and Norris derivative(ND)，partial least square regression(PLS)was used to establish near-infrared spectrometry quantitative models for the determination of asiaticoside.Results：In validation set，the relative error(RSEP)，and the ratio of standard deviation and standard error of prediction were 0.0352 and 6.47，respectively，which indicated that the models had good accuracy.The root mean square error （RMSEC）in calibration set and the root mean square error(RMSEP)in validation set were 1.17 and 1.25，respectively，and the root mean square error of cross validation(RMSECV)was 3.43，which showed that the near infrared model was robust.Conclusion：The established models have good accuracy and robustness，which provide an efficient approach for the fast determination of asiaticoside in Centella Total Glucosides.