Abstract: For rapidly determining the moisture retention of tobacco leaf, a near infrared(NIR)model for moisture retention of tobacco leaf was developed by using equilibrium moisture desorption rate of tobacco leaf as an evaluation index. The temperature and humidity range for determining the moisture retention were 22℃ and 60%-30%, respectively. The results showed that:1) The optimal spectral pretreatment method for the model was vector normalization, and the optimal wavelength range was 7 502.0-4 246.7 cm^-1. The RMSECV and correlation coefficient of the model were 0.256 and 0.928 1, respectively. 2) The external validation of the model was conducted, the average absolute deviation and the average relative deviation of true values and predicted values were 0.15 percentage point and 1.80%, respectively. This model can quickly and effectively measure the moisture retention of tobacco leaf with good prediction effect.