Site index is an important characteristic for forest management and growth modelling, especially with regard to the ongoing environmental changes. This study introduces a new methodological approach to detecting and quantifying long-term changes in site index. It consists of: (1) a module to construct a reference site index curve from the data under investigation, (2) an OLS-based-CUSUM test to identify change points in development of site index over time, (3) a composite estimator for quantifying changes of the site index over time. The approach was developed using data from 541 Norway spruce (Picea abies [L.] Karst.) long-term experimental plots in southwest Germany between 1872 and 2010. The method is designed to exploit height measurement series of stands growing at different times and different locations (temporally and spatially disjunct). The approach indicated a change in the trajectory of site indices in the early 1950s. Site indices were stable until the change point and have displayed an almost continuous increase since then.