This article presents the development and application of computer-simulation methods to predict stress-related adaptive bone remodeling, in accordance with 'Wolff's Law'. The models are based on the Finite Element Method (FEM) combined with numerical formulations of adaptive bone-remodeling theories. The Strain Energy Density (SED) is used as a feedback control variable to determine shape or bone density adaptations to alternative functional requirements, assuming homeostatic SED distribution as the remodeling objective. These models are applied to investigate the relationship between 'stress shielding' and bone resorption in the femoral cortex around intramedullary prostheses, such as used in Total Hip Arthroplasty (THA). The study shows that the amount of bone resorption depends mainly on the rigidity and bonding characteristics of the implant. Homeostatic SED can be achieved when the resorption process occurs at the periosteal surface rather than inside the cortex, provided that the stem is adequately flexible. The results provide quantitative hypotheses about the predicted effects of stem rigidity and degree of fixation, which are compatible with animal experimental data. The study also predicts differences in the adaptive remodeling process between loose and well-fixed prostheses.This article presents the development and application of computer-simulation methods to predict stress-related adaptive bone remodeling, in accordance with 'Wolff's Law'. The models are based on the Finite Element Method (FEM) combined with numerical formulations of adaptive bone-remodeling theories. The Strain Energy Density (SED) is used as a feedback control variable to determine shape or bone density adaptations to alternative functional requirements, assuming homeostatic SED distribution as the remodeling objective. These models are applied to investigate the relationship between 'stress shielding' and bone resorption in the femoral cortex around intramedullary prostheses, such as used in Total Hip Arthroplasty (THA). The study shows that the amount of bone resorption depends mainly on the rigidity and bonding characteristics of the implant. Homeostatic SED can be achieved when the resorption process occurs at the periosteal surface rather than inside the cortex, provided that the stem is adequately flexible. The results provide quantitative hypotheses about the predicted effects of stem rigidity and degree of fixation, which are compatible with animal experimental data. The study also predicts differences in the adaptive remodeling process between loose and well-fixed prostheses.