Abstract: In this Thesis, we derive a theoretical Optical Potential (OP) from the Taylor expansion of the mass operator of the one-particle Green function, which is obtained from a G-matrix Brueckner-Hartre-Fock calculation using the effective Gogny force instead of a microscopic interaction. In nuclear matter the real part of this OP is provided by the first-order term of the Taylor expansion, which coincides with the mean-field Hartree-Fock potential felt by the projectile. The imaginary part is extracted from the second-order term of the Taylor expansion and corresponds to uncorrelated 2p-1h states in the intermediate states assuming only two-body interactions. The OP potential in finite nuclei is thus obtained through the Local Density Approximation using Hartree-Fock wave functions of the target nucleus computed with the same Gogny force. This Gogny-based OP predicts a reasonable agreement between different scattering observables (cross-sections and analyzing powers) and the experimental data. A more quantitative agreement, like to that found using the Koning-Delaroche OP, can be obtained by a suitable renormalization of the real and imaginary parts of the Gogny-based OP. In the second part, we have built another OP for describing the elastic scattering of light particles, such as 2H, 2H, 3He, and 4He, with a generalized Watanabe model, which uses the Gogny-based nucleon-nucleus OP and the ground-state wave functions. This model reproduces the experimental data rather well for low scattering angles but fails for large angles. This fact points out that more sophisticated calculations, such as the Continuum Discretized Coupled Channel or the Coupled Reaction Channel, on top of our results are needed to improve the agreement with the experiment. Finally, we study quasi-elastic (p,n) reactions within the Lane model, where the exact isospin symmetry is assumed. The calculation is carried out using the Distorted Wave Born Approximation together with Gogny nucleon-nucleus OP results. This model provides a reasonably good agreement with the experimental data for different reactions spanned along the periodic table at different energies.