The effects of exercise on stress fracture risk are paradoxical. Exercise can promote both bone formation and resorption, which in turn, can reduce and increase risk of stress fractures, respectively. We review classic and current literature that suggests that the processes that underlie these responses to exercise are distinct. Bone remodeling involves osteoclastic resorption of fatigue-damaged bone, coupled with subsequent bone deposition to replace the damaged tissue. Bone modeling involves the independent action of osteoblasts and osteoclasts forming or resorbing bone, respectively, on a surface. In the formation mode, modeling results in increased bone stiffness, strength, and resistance to fatigue. Both the remodeling and modeling responses to exercise require significant time for newly deposited bone to fully mineralize. We propose that recognizing these two distinct physiologic pathways and their related time courses reveals the theoretical basis to guide exercise prescription to promote bone health during periods of heightened stress fracture risk. Such guidance may include minimizing rapid increases in the duration of repetitive exercises that may cause fatigue damage accrual, such as long-distance running and marching. Rather, limiting initial exercise characteristics to those known to stimulate bone formation, such as short-duration, moderate-to-high impact, dynamic, and multidirectional activities with rest insertion, may increase the fatigue resistance of bone and consequently minimize stress fracture risk.