Purpose The presence of bone metastases has excluded participation of cancer patients in exercise interventions and is a relative contraindication to supervised exercise in the community setting because of concerns of fragility fracture. We examined the efficacy and safety of a modular multimodal exercise program in prostate cancer patients with bone metastases. Methods Between 2012 and 2015, 57 prostate cancer patients (70.0 ± 8.4 yr; body mass index, 28.7 ± 4.0 kg·m -2) with bone metastases (pelvis, 75.4%; femur, 40.4%; rib/thoracic spine, 66.7%; lumbar spine, 43.9%; humerus, 24.6%; other sites, 70.2%) were randomized to multimodal supervised aerobic, resistance, and flexibility exercises undertaken thrice weekly (EX; n = 28) or usual care (CON; n = 29) for 3 months. Physical function subscale of the Medical Outcomes Study Short-Form 36 was the primary end point as an indicator of patient-rated physical functioning. Secondary end points included objective measures of physical function, lower body muscle strength, body composition, and fatigue. Safety was assessed by recording the incidence and severity of any adverse events, skeletal complications, and bone pain throughout the intervention. Results There was a significant difference between groups for self-reported physical functioning (3.2 points; 95% confidence interval, 0.4-6.0 points; P = 0.028) and lower body muscle strength (6.6 kg; 95% confidence interval, 0.6-12.7; P = 0.033) at 3 months favoring EX. However, there was no difference between groups for lean mass (P = 0.584), fat mass (P = 0.598), or fatigue (P = 0.964). There were no exercise-related adverse events or skeletal fractures and no differences in bone pain between EX and CON (P = 0.507). Conclusions Multimodal modular exercise in prostate cancer patients with bone metastases led to self-reported improvements in physical function and objectively measured lower body muscle strength with no skeletal complications or increased bone pain.