Abstract: In order to realize rapid prediction of fresh walnut kernel moisture during shelf period，hyperspectral technology was used to collect spectral data of walnut green peel and moisture content of walnut kernel was measured.Eleven characteristic wavelengths were extracted by continuous projection method（SPA）to establish partial least squares（PLS），multiple linear regression（MLR），least squares-support vector machine（LS-SVM）model.Results showed that the LS-SVM modeling effect was the best，correlation coefficient RP of the prediction set was 0.800 7，and root mean square error RRMSEP was 2.189 7.Studies have shown that the use of hyperspectral technology could achieve rapid prediction of moisture of fresh peel walnut kernels during shelf life，providing guidance for storage，transportation and marketing of fresh peel walnuts.