To achieve the rapid detection of key indicators of Jiang-flavor baijiu (Chinese spirit) base liquor, analytical models of seven key components in Jiang-flavor baijiu were established using near infrared (NIR) spectroscopy combined with a partial least squares (PLS) algorithm, with gas chromatography and titration used as reference methods. Characteristic bands of various substances were screened by the interval partial least squares (iPLS) method, and the optimal spectral preprocessing method was determined from twelve individual or combined approaches. The coefficients of determination of cross-validation (R2Val) of optimized models for acetic acid, propionic acid, ethyl acetate, ethyl lactate, furfural, isoamyl acetate, and total acid were in the range of 0.854 2~0.963 8, and the root mean square errors of cross-validation (RMSECV) were in the range of 0.003 8~0.515 8. The standard errors of prediction (SEP) for external validation were in the range of 0.002 8~0.478 5, and residual predictive deviations (RPD) were in the range of 2.41~6.43, indicating good model prediction accuracy and robustness. In summary, NIR spectroscopy achieved rapid detection of the acetic acid, propionic acid, ethyl acetate, ethyl lactate, furfural, isoamyl acetate, and total acid contents in Jiang-flavor baijiu base liquor, improving the detection efficiency considerably and enabling rapid feedback in the production process. The results of this study support the application of online NIR spectroscopy in intelligent brewing.