Introduction This study aimed to examine the efect of high-load resistance training (HLRT) on bone mineral density (BMD) in patients with osteoporosis and osteopenia using a meta-analysis. Materials and methods We searched for randomized controlled trials (RCTs) on HLRT in patients with osteoporosis and osteopenia from medical databases. Our meta-analysis was performed with the primary endpoints being the standardized mean diference (SMD) of the change in BMD of the lumbar spine (LS), femoral neck (FN), and total hip (TH). The robustness of the results was assessed by subgroup analysis. Heterogeneity factors were examined by meta-regression. Publication bias was evaluated using a funnel plot. Results We selected nine RCTs, with 259 patients in the HLRT group (women, 55.2%) and 236 patients in the control group (women, 62.7%). The HLRT group showed a signifcant increase in BMD in the LS [SMD=1.40, 95% confdence interval (CI)=0.68–2.12, p<0.001, I 2=90%], the FN (SMD=0.86, 95% CI=0.05–1.67, p=0.04, I 2=92%), and the TH (SMD=1.26, 95% CI=0.45–2.08, p=0.002, I 2=91%). Subgroup analysis confrmed the robustness of the results only in LS. Total sessions and a high risk of bias were identifed as the factors of heterogeneity in FN and TH (p<0.05). The funnel plot showed asymmetry in all measurement sites. Conclusion This study suggested that HLRT can be efective in increasing BMD, mainly of LS, in patients with osteoporosis and osteopenia. However, due to high heterogeneity and publication bias, additional studies with a low risk of bias should be conducted to generalize our fndings.