With the progress of science and technology, society is evolving toward intelligence. To use real-time paper literature monitoring to realize the preservation of paper literature will be an inevitable development trend in the future. In order to provide the basis for the real-time monitoring of library literature, Fourier transform medium infrared photo-acoustic spectroscopy is used to determine the water content in paper in combination with stoichiometry. A quantitative model of paper water content is established by Principal Component Regression (PCR) and Partial Least Squares Regression (PLSR) respectively on the basis of medium-infrared photoacoustic spectroscopy. The optimum primary component number is selected by cross-validation. The result shows that the modeling result of partial least squares regression is better than that of the principal component regression (the determination coefficient is 0.3681>0.3532). By increasing the number of principal component, the prediction of the models will be better. However, there are some risks of fitting. In the future, more paper samples will be collected so as to establish a stable paper quality detection model and lay a basis for the real-time infrared spectrum monitoring of paper literature.