As usually known, the physics in space have a tendency to behave differently from the physics on earth. In fact, Free Molecular Flow are an important hurdle to cross in order to completely gasp molecules behaviors. Outgassing for example is one of these phenomena that deeply impacts satellite performances and remains complex to model regardless of the numerous successful optical satellites missions. Despite the efforts spent last years, on-ground outgassing testing still gives little information about the driving mechanism of outgassing. Is it controlled by desorption? By diffusion? Is it a mix of both? And once extracted from the material bulk, what are these molecules? Are they carbon based molecules? Are they a mix of different species? And once they leave the materials, how the molecules condensate on the sensitive surface? How are they reemitted? Is it governed by reflection or by a residence time? In addition, the configuration on a spacecraft is far more complicated than in a simple set-up: huge temperature gradients exists, a material with low outgassing criteria (TML, RML, CVCM) can still be a threat to the whole mission if it is located next to a sensitive optical surface… Meanwhile, OHB conducted several numerical investigations that this article will summarize. Based on latest JAXA1 and ONERA² experiences, a diffusion model have been implemented and connected to the view factors calculations of a closed cavity. The model is based on the Fick’s law and a numerical calculation was performed thanks to an Euler/Crank-Nicolson method. Compared to the desorption modelling, the diffusion results showed quite important differences especially in the starting of the outgassing. To able to judge between the two approaches, the numerical model have been confronted to an outgassing set-up called the “CNES mock-up”. To make things more complicated, the “reemission” part is also challenging. Depending on the configuration to model, reflection of molecules can be the driving mechanism and the residence time approach is too slow to catch up with the real observed behavior. Concrete examples will be presented in this article to open the discussion on possible approaches to better model and understand outgassing.