To evaluate the applicability of non-invasive near infrared reflectance (NIR) spectroscopy for detection of bottled vinegar, transmission spectra of bottled vinegar samples were obtained and compared with the spectra obtained by using cell detection method, where samples were titrated in the fixed liquid cell with a 1.0 mm light path length. The result shows that, the spectra obtained by the two different methods have obvious differences in the near infrared region. Firstly, the spectra obtained by using bottled detection method have no absorption peak at 1 480 and 1 900 nm, and furthermore, it has absorption peak at 1 660 nm. Secondly, the max. absorbance value of bottled detection is less than 4, while with the cell detection it is near 6. It indicates that glass packaging has influence on the detection of bottled vinegar. The 1st derivative method was put forward to eliminate this influence, and then qualitative analysis model was obtained by using principal component analysis-artificial neural net work. The precision of prediction result achieved 100%. This research shows that 1st derivative can eliminate the influence of glass packaging, and realize the qualitative analysis of bottled vinegar.