Résumé
We consider the production, at future lepton colliders, of charged Higgs pairs in supersymmetric models. Assuming a relatively light SUSY scenario, and working in the MSSM, we show that, for c.m. energies in the one TeV range, a one-loop logarithmic Sudakov expansion that includes an "effective" next-to subleading order term is adequate to the expected level of experimental accuracy. We consider then the coefficient of the linear (subleading) SUSY Sudakov logarithm and the SUSY next to subleading term of the expansion and show that their dependence on the supersymmetric parameters of the model is drastically different. In particular the coefficient of the SUSY logarithm is only dependent on $\tan\beta$ while the next to subleading term depends on a larger set of SUSY parameters. This would allow to extract from the data separate informations and tests of the model.