This article presents a method for identifying the socio-economic determinants of transformations in agroforestry systems (AFS). It is based on the analysis of diachronic data collected on three levels: country, region and farm. It offers a new method of spatio-temporal triangulation, making it possible to identify and classify the main socio-economic determinants of these transformations. It proposes an evaluation grid based on seven categories of socio-economic determinants. These inputs are then used to produce models, which explain the logic of the transformations in AFS and the socio-economic determinants involved. Examples of Cameroonian cocoa-based agroforestry systems serve to illustrate the main steps and modalities for using this method and the results it can provide.