The Effective Fragment Potential (EFP) is a quantum-mechanical based model used to describe non-covalent interactions of small molecules or fragments. It can be used along with fully <i>ab initio</i> methods to study the electronic properties of complex systems, such as solvated chromophores or proteins. For this purpose, the system is divided into two regions: one modeled with quantum mechanics and the other with EFP. The interaction between the QM region and the effective fragments has popularly been described through electrostatics and polarization only. This thesis focuses on the development of the QM/EFP exchange-repulsion term, as well as the evaluation of the dispersion term and a charge-penetration correction. The goal of is to determine how these terms can increase the accuracy of QM/EFP calculations without an increase in their computational cost.