The demo-genetic model (Coulson et al, 2006; Benton et al, 2006) is generally used in population dynamics and evolutionary studies. In this very prospective presentation, we will question about the possible use of this approach in the fields of long term breeding. The CAPSIS platform (http://capsis.cirad.fr/) was initially developped for modelling stand growth and silviculture (de Coligny et al, 2004). genetics features were then added to this software (Dreyfus et al, 2005). It is not directly a simulation tool but a modelling platform well suited for the implementation of one's own demo-genetic models and for the development of ad-hoc simulation tools. The flexibility of this modelling platform allows to combine biological processes of various kinds (demography, genetics, physiology, biotic interactions...) together with sylvicultural treatments, in an individual-based and spatially explicit way. The challenge is then to have relevant models of these processes, which can be combined using the tools and libraries available in the modelling platform. After a very brief survey of the modules that currently use the Genetics library, in trees and fishes, we will investigate possible applications in two main directions of interest for breeding: questions related to spatial processes and questions related to the make-up of fitness or breeding value. The approach could be used with very diverse objectives, such as: test the genetic consequences of gene flow in a composite population after natural regeneration under various spatial organisations, simulate long term stand growth for sensitivity analysis to choose the best selection strategy, predict the Qst/Fst ratio (i.e. local adaptation vs fixation of QTLs) in various scenarios, simulate datasets to test marker-based methods of estimation of genetic parameters, predict changes in genetic parameters at different stages of stand development