Gravity retaining wall is an essential structure widely used in civil engineering, roadway, highway construction, breakwater, and coastal protection. However, the dimensions of this structure are still determined based on experience and safety, leading to a surplus of building materials. This paper presents the optimization of gravity retaining walls using Swarm Intelligence algorithms. This research studies the case of a stone masonry gravity retaining wall. A complex step cross-section wall laid on a stone base is proposed. The objective function is based on the cost proportion of the stone masonry part and the stone base, while the constraints are chosen based on the geometry constraints of the structure, the bearing capacity of the materials, and the overall stability of the retaining wall as well as the foundation. Especially the role of the stone base is first time considered.