Background The data of head-to-head comparisons of the effect of bone-modifying agents (BMAs) in patients with androgen deprivation therapy (ADT) for prostate cancer without skeletal metastasis is limited. Thus, we conducted a network meta-analysis to compare each BMA for the efficacy of bone mineral densities (BMDs) and the risk of fracture. Methods We performed a network meta-analysis to compare the change of BMDs and the risk of vertebral fracture in the studies included using a random-effect model. The primary outcomes are the change of BMD of the lumbar spine (LS) and the total hip (TH) from the baseline at 1 year from the initiation of the BMA and the risk of vertebral fracture. Results We identified and included 15 studies in this analysis. All BMAs except risedronate showed a significant increase of BMD of the LS compared with groups without BMA, among which zoledronate showed the most BMD gain. At TH, bisphosphonates (alendronate, pamidronate, and zoledronate) and denosumab showed significant elevation compared with the no-BMA group. Denosumab was associated with the most BMD gain at the TH. Only denosumab reduced the risk of vertebral fracture (relative risk 95% confidence interval: 0.40 0.20–0.81). Although zoledronate showed the highest BMD gain at the LS, it did not reduce the risk of vertebral fracture in this analysis. Conclusion Most bisphosphonates and denosumab significantly increased BMD at the LS and the TH in patients receiving ADT for prostate cancer without skeletal metastasis. In particular, zoledronate and denosumab were the most potent BMAs in terms of BMD increment at the LS and the TH, respectively. However, denosumab, not zoledronate, was the only BMA that showed a significant risk reduction of vertebral fracture. We need further studies to examine the change of bone quality and the effect on the risk of non-vertebral and hip fractures.