GHOUL, RayanDEGAICHI, Nouar2026-10-202026-06-02https://dspace.univ-tebessa.dz/handle/123456789/293This work presents a study of the Conjugate Gradient method for solving large, sparse systems of linear equations. The method relies on constructing a sequence of conjugate directions and determining optimal step sizes to efficiently minimize the associated quadratic function. Its main numerical properties, including convergence behavior and computational efficiency, are examined and discussed.enConjugate Gradientsparse systemslinear equationsconvergencenumerical meth- ods.Thesis