Alzheimer’s disease manifests itself in brain tissue by neuronal death, due to aggregation of β-amyloid, produced by senile plaques, and hyperphosphorylation of the tau protein, which produces neurofibrillary tangles. One of the genetic markers of the disease is the gene that translates the presenilin-2 protein, which has mutations that favor the appearance of the disease and has no reported crystallographic structure. In view of this, protein modeling is performed using prediction and structural refinement tools followed by an energetic and stereochemical characterization for its validation. For the simulation, four reported mutations are chosen, which are Met239Ile, Met239Val, Ser130Leu, and Thr122Arg, all associated with various functional responses. From a theoretical analysis, a preliminary bioinformatic study is made to find the phosphorylation patterns in the protein and the hydropathic index according to the polarity and chemical environment. Molecular visualization was carried out with the Chimera 1.14 software, and the theoretical calculation with the hybrid quantum mechanics/molecular mechanics system from the semi-empirical method, with Spartan18 software and an AustinModel1 basis. These relationships allow for studying the system from a structural approach with the determination of small distance changes, potential surfaces, electrostatic maps, and angle changes, which favor the comparison between wild-type and mutant systems. With the results obtained, it is expected to complement experimental data reported in the literature from models that would allow us to understand the effects of the selected mutations.