Patients with nonalcoholic fatty liver disease (NAFLD) are reported to have a higher risk of osteoporosis/fractures; however, the causal relationship remains unclear. Publicly available genome-wide association studies (GWASs) were used for Mendelian randomization (MR) analysis. GWASs of NAFLD and fractures were obtained from the FinnGen Consortium. GWASs of bone mineral density (BMD) were derived from a meta-analysis. GWASs of obesity, diabetes, liver function, and serum lipid-related metrics were used to clarify whether the accompanying NAFLD symptoms contributed to fractures. Moreover, two additional GWASs of NAFLD were applied. A causal association was not observed between NAFLD and BMD using GWASs from the FinnGen Consortium. However, a causal relationship between NAFLD and femoral neck-BMD (FN-BMD), a suggestive relationship between fibrosis and FN-BMD, and between NAFLD and osteoporosis were identified in replication GWASs. Genetically proxied body mass index (BMI), high-density lipoprotein (HDL), and hip circumference increased the likelihood of lower limb fractures. The waist-to-hip ratio decreased, whereas glycated hemoglobin (HbA1C) and homeostasis model assessment of β-cell function (HOMA-B) increased the risk of forearm fractures. Low-density lipoprotein (LDL) reduced, whereas HbA1C increased the incidence of femoral fractures. Alkaline phosphatase (ALP) raised the risk of foot fractures. However, after a multivariate MR analysis (adjusted for BMI), all the relationships became insignificant. NAFLD caused reduced BMD, and genetically predicted HDL, LDL, HbA1C, HOMA-B, ALP, hip circumference, and waist-to-hip ratio causally increased the risk of fractures. BMI may mediate causal relationships. Larger GWASs are required to verify this finding.