Near-infrared spectroscopy is a spectral analysis technique that involves electromagnetic radiation at wavelengths between the visible spectrum and the mid-infrared spectrum. The technique can realize non-destructive testing of samples and is particularly suitable for rapid analysis in the field and online quality monitoring. This technique utilizes the absorption of near-infrared light by hydrogen-containing groups within a sample, enabling the extraction of specific information regarding the characteristics of these hydrogen-containing groups within the analyzed organic molecules. The application of chemometrics in near-infrared spectral analysis reduces the data set size, excludes noise interference and improves the efficiency of multivariate correction. The finite variable model construction reduces the random errors and improves the accuracy of the prediction results. The present document synthesizes the application of near-infrared spectroscopy and chemometrics in the context of liquid food safety detection. It provides a comprehensive overview of the state-of-the-art research in this field, emphasizing the role of chemometrics analysis in evaluating the quality and safety of liquid foods. Additionally, it delves into the merits and limitations of this technological approach, offering insight into its potential for future advancement and expansion.