A near infrared diffuse spectroscopy technology combined with optical fiber sensing technology is used to explore nondestructive testing methods to identify the picking period of juicy peaches. 48 each of juicy peaches of 3

2 and 1 days from the best picking period and the day in the best picking period from a greenhouse in Yangshan town of Wuxi. The samples are collected by near infrared spectrometer. Three pre-processing methods i. e. smoothing

first derivative and mutiplicative scatter correction are used. Identification models of picking period are developed by principal component analysis（ PCA） and partial least square（ PLS） regression. The results show that first derivative and smoothing combination preprocessing construct the best predicted model. The correlation coefficient of calibration and validation model are 0. 927 9 and 0. 913 8 respectively. The root mean square error of cross validation（ RMSECV） and the root mean square error of prediction（ RMSEP） are 0. 3003 and 0. 3349 respectively. The recognition rates in the calibration set and prediction set of juicy peaches are 95. 13 % and93. 75 % respectively. Results shows that the method of using the near infrared spectroscopy technology to identify the picking period of juicy peaches has a very good application prospect.