The excipients contents in the Amlodipine Besylate Tablet were rapidly detected combing of the near infrared spectroscopy and Chemometrics. Characteristic wavelength variation points were screened by methods of Random Frog, Variable Importance Projection and Competition Self-Adaptive Reweighted Sampling. After processing the original spectrum by 9 kinds of spectrum pre-processing methods, the Partial Least Squares (PLS) model and Support Vector Regression analysis (SVR) model were established respectively and compared to each other. And then the optimized model was applied to test the samples. The results show that the effect of Random Frog is better for the selection of optimal characteristic wavelength variables in the samples involved; For the model predictions, the effect of PLS quantitative model is better for the evaluation parameters in the determination coefficient and RMSEP of the five indexes, when compared with that of the SVR model, and the Relative Percent Difference (RPD) values are all more than 3.0. The standard error of the tested values and measured values for samples are both less than 1.30; Paired t-test shows that there is no significant difference at the significance level of a=0.05. So near infrared diffuse reflectance spectroscopy can be used to rapidly detect the excipients′ contents in the Amlodipine Besylate Tablets, with a good repeatability, an intermediate precision, a linearity, accuracy, can provide a good reference for the rapid detection of other pharmaceutical excipients′contents.