This study proposes a numerical methodology to minimize the bone mass loss in a femur with a total hip arthroplasty procedure, considering uncertainties in the material parameters and using a reliability-based design optimization (RBDO) procedure. A genetic algorithm (GA) is applied for optimization, and a three-dimensional finite element (FE) model associated with the bone remodeling procedure is proposed and described to account for the internal and external femoral bone behavior. An example of a femoral prosthesis design is presented as a basis for discussion of the proposed methodologies, and the corresponding reliability level is evaluated. Constraints on the strength of all materials and target reliability levels are inputs to the optimization model. The main prosthesis dimensions and Young modulus are the design variables. The proposed methodology is compared with a well-known deterministic optimization (DO) procedure and the results show that it is important to consider the uncertainties in this kind of problem since in this case, the a posteriori reliability may be low.