The ionization and excitation of atoms and molecules in the upper atmospheres of the Earth and planets are computed by a number of physical models. From these calculations, quantities measurable by dedicated satellite experiments such as airglow and electron fluxes can be derived. It is then possible to compare model and observation to derive more fundamental physical properties of the upper atmospheres, for example, the density as a function of altitude. To ensure the accuracy of these retrieval techniques, it is important to have an estimation of the uncertainty of these models and to have ways to account for these uncertainties. The complexity of kinetic models for computing the secondary production of excited state species (including ions) makes it a difficult evaluation, and studies usually neglect or underestimate it. We present here a Monte-Carlo approach to the computation of model uncertainties. As an example, we studied several aspects of the model uncertainties in the upper atmosphere of Mars, including the computed secondary electron flux and the production of the main ion species. Our simulations show the importance of improving solar flux models, especially on the energy binning and on the photon impact cross sections, which are the main sources of uncertainties on the dayside. The risk of modifying cross sections on the basis of aeronomical observations is highlighted for the case of Mars, while accurate uncertainties are shown to be crucial for the interpretation of data from the particle detectors onboard Mars Global Surveyor. Finally, it shows the importance of AtMoCiad, a public database dedicated to the evaluation of aeronomy cross section uncertainties. A detailed study of the resulting emissions cross sections uncertainties is the focus of a forthcoming paper (Gronoff et al., 2012) in which the outputs discussed in the present paper are used to compute airglow uncertainty, and the overall result is compared with the data from the SPICAM UV spectrometer onboard Mars Express.