CropMetaPop is a new simulation tool to model the genetic evolution of crop diversity under on-farm dynamic management. Under this type of conservation and use of varieties, seeds are resown and exchanged between farmers and the set of connected populations is described as a crop metapopulation. CropMetaPop is therefore at the interface of genetic and social processes. We used sensitivity analyses to check the behaviour of the model and to identify which parameters and range of values for them induce the most variability in the outputs. CropMetaPop was found to behave as expected. Depending on the type of locus studied (neutral or selected), the parameters related to drift or selection were those that induced the most variability in the outputs. Colonisation, migration, and network topology parameters were less influential. Looking at the detailed results will help setting the parameters to relevant values in the future utilisation of the model.