Abstract: 【Objective】 Real-time and rapid acquisition of crop moisture status is extremely important for accurate management of crop water in farmland. 【Methods】 This study took typical drip irrigation winter wheat in Xinjiang as the research object, applied hyperspectral imaging technology to obtain winter wheat canopy spectrum information, and smoothed and conversed the data of the raw spectral reflectance. Simple linear regression (SLR), principal components regression (PCR) and partial least squares regression (PLSR) methods were used to construct plant moisture content estimation models for the raw spectrum and transformed spectrum of winter wheat canopy respectively. 【Results】 The results showed that the correlation between the raw spectral reflectance of winter wheat canopy and the plant moisture content was not high. Data transformation of the raw spectral reflectance significantly enhanced the correlation with the moisture content and the number of relevant bands. Among them, the correlation coefficient between reciprocal first-order differential transformation and winter wheat plant moisture content was the largest, which was -0.893,0. However, the band position corresponding to the maximum correlation coefficient of each transformation was not fixed. Among the three modeling methods, the PLSR method had the highest model accuracy, while the logarithmic transformed PLSR model had the highest estimation accuracy. The model R_p^2, RMSE_p and RPD values were 0.880,8, 3.251,2%, and 2.934,3, respectively. There were differences in the estimation model accuracy of winter wheat at different growth stages. The estimation model accuracy of the jointing stage and the heading stage was lower, and the mid-grouting stage was the highest. The estimated model R_p^2, RMSE_p and RPD values were 0.904,8, 1.381,1%, 3.454,7, respectively. 【Conclusion】 It is feasible to use hyperspectral imaging technology to estimate the moisture content of winter wheat plants, and the best estimation effect is in the mid-grouting stage.