Abstract Background Low-density lipoprotein cholesterol (LDL-C) is suggested to play a role in osteoporosis but its association with bone metabolism remains unclear. Effects of LDL-C-lowering drugs on bone are also controversial. We aim to determine whether LDL-C is linked causally to bone mineral density (BMD) and assess the effects of LDL-C-lowering drugs on BMD. Methods Association between blood lipid levels and BMD was examined by epidemiological observation analyses in a US representative cohort NHANES III (n = 3638) and the Hong Kong Osteoporosis Study (HKOS; n = 1128). Two-sample Mendelian randomization (MR), employing genetic data from a large-scale genome-wide association study (GWAS) of blood lipids (n = 188 577), total body BMD (TB-BMD) (n = 66 628) and estimated BMD (eBMD) (n= 142 487), was performed to infer causality between LDL-C and BMD. Genetic proxies for LDL-C-lowering drugs were used to examine the drugs’ effects on BMD. Results In the NHANES III cohort, each standard deviation (SD) decrease in LDL-C was associated with a 0.045 SD increase in femoral neck BMD (95% CI: 0.009 − 0.081; P = 0.015). A similar increase in BMD was observed in the HKOS at femoral neck and lumbar spine. In MR analysis, a decrease in genetically predicted LDL-C was associated with an increase in TB-BMD {estimate per SD decrease, 0.038 [95% confidence interval (CI): 0.002 − 0.074]; P = 0.038} and eBMD [0.076 (0.042 − 0.111); P = 1.20x10−5]. Reduction in TB-BMD was causally associated with increased LDL-C [0.035 (0.033 − 0.066); P = 0.034]. Statins’ LDL-C-lowering proxies were associated with increased TB-BMD [0.18 (0.044 − 0.316); P = 9.600x10−3] and eBMD [0.143 (0.062 − 0.223); P = 5.165x10−4]. Conclusions Negative causal association exists between LDL-C level and BMD. Statins’ LDL-C-lowering effect increases BMD, suggesting their protective effect on bone.