An innovative survey was conducted of the collections of the historical Biblioteca Classense, located in the urban area of Ravenna (Northern Italy). The survey aimed to evaluate the current conservation state of the book collections, where 297 paper-based items, including incunabula, manuscripts and books, dating from the 14th to the 20th century, were selected for analysis. This innovative survey was carried out non-destructively by assessing degradation visually and by measuring NIR spectral data followed by multivariate data analysis. Chemical and physical paper properties, important for paper characterisation and implementation of conservation strategies were determined, including paper type, pH, degree of polymerisation (DP), tensile strength, lignin, protein, and rosin content. This survey provided a significant quantitative dataset for rag paper covering a 600-year period. The analysis of DP changes over time allowed the first experimental estimation of the rate constant for historical rag paper, i.e., (4.2 ± 0.6)•10 −7 year −1 , which was validated with predictions based on the Collections Demography dose response function for historic paper taking into account the past climate in Ravenna. Statistical methods were employed to describe the correlations between the measured variables and different features of the books, suggesting that the degree of polymerisation can be used as a general proxy for rag paper mechanical strength.