Objective: To develop a method for the determination of ursolic acid content in Eriobotrya japonica by near-infrared spectroscopy. Method: The ursolic acid contents of 131 samples were determined by the method of HPLC. NIR spectrograms were measured

the second derivative was used as spectral preprocessing options

and calibration model of the ethanol extracts content was established by the partial least squares regression analysis

The models were verified by internal cross validation and external predictive validation with 25 E. japonica samples. Result: The correlation coefficients

the root-mean-square error of calibration

the root-mean-square error of prediction and the toot-mean-square error of cross-validation of the calibration model for ethanol extracts content were 0.982 46

0.175

0.249 and 0.525 02. Conclusion: The method is steady

accurate

and can be used to predict ethanol extracts content of Eriobotrya japonica rapidly.