In order to develop a method for rapid detection of moisture content in vinegar residue substrate, the spectral data of 69 representative samples were collected by using visible-near-infrared spectroradiometer, and the oven-drying method was applied to obtain the data of these samples moisture content. Partial least squares regression (PLS) method was selected to perform the calibration models for predicting the moisture content. For determining the number of principal component factors in the PLS model and the method of spectra preprocessing, the model with the lowest prediction residual error sum of square based on cross-validation for the calibration samples set was chosen. The determination coefficient between the predicted and the reference results for the prediction samples set, along with the root mean squared error of prediction (RMSEP), was used as evaluation parameters for the models. The results indicate that the forecast result of spectral data is optimal through the spectra preprocessing method of moving average filter (MAF) and first derivative (FD), with 5 principal component factors. The determination coefficients of the calibration and prediction models are 0.9930 and 0.9901, with root mean squared error of calibration (RMSEC) and prediction (RMSEP) of 0.0676 and 0.0715, respectively. Therefore， visible-near-infrared spectroscopy with PLS method can be used successfully to determine the moisture content of vinegar residue substrates.