The method of near infrared reflectance spectroscopy for predicting volatile oil content in intact Zanthoxylum bungeagum Maxim was developed. Adopting partial least squares (PLS) regression, the prediction models were built by the 118 samples in 9 kinds acquisition parameters combination with the resolution (4, 8 and 16cm-1) and the sample scans (32, 64 and 128). Results demonstrated that the model obtained at resolution of 16 cm-1and sample scans of 128 was better than the others, and the determination coefficient and RMSECV of cross validation were 0.907 and 0.509, respectively. Applying the model to the test set with 20 samples, the determination coefficient, RMSEP and RPD of test set validation were 0.973, 0.272 and 6.28, respectively. The experimental results show that NIRS can be used as a method to detect intact Zanthoxylum bungeagum Maxim volatile oil content rapidly and nondestructively. 