This case-control study investigated the feasibility of street football as a health-enhancing activity for homeless men, specifically the musculoskeletal effects of 12 weeks of training. Twenty-two homeless men participated in the football group (FG) and 10 served as controls (C). Plasma osteocalcin, TRACP5b, leptin, and postural balance were measured, and whole-body DXA scanning was performed. The attendance rate was 75% (2.2 ± 0.7 sessions per week). During 60 min of training, the total distance covered was 5534 ± 610 m, with 1040 ± 353, 2744 ± 671, and 864 ± 224 m covered by high-intensity, low-intensity, and backwards/sideways running, respectively. In FG, osteocalcin increased by 27% from 20.1 ± 11.1 to 25.6 ± 11.8 ng/mL (P = 0.007). Postural balance increased by 39% (P = 0.004) and 46% (P = 0.006) in right and left leg. Trunk bone mineral density increased by 1.0% from 0.959 ± 0.095 to 0.969 ± 0.090 g/cm(2) (P = 0.02). No effects were observed in C. In conclusion, street football appears to be a feasible training activity with musculoskeletal health benefits for homeless men. The attendance rate and the training intensity were high, and 12 weeks of training resulted in a substantial anabolic response in bone metabolism. Postural balance improved markedly, and the overall risk of falling, and hospitalization due to sudden trauma, could be reduced by street football for homeless men.