Observational studies indicate that sleep apnea is associated with non-alcoholic fatty liver disease (NAFLD) and its related metabolic features, independent of confounding factors including obesity. However, the causal relationships remain to be determined. Univariable and multivariable Mendelian randomization (MR) analyses were performed to investigate the causal relationship between sleep apnea and NAFLD, along with its typical features including liver function, glycemic traits and lipid profiles. Summary-level data for sleep apnea were obtained from the Finngen consortium (33,423 cases and 307,648 controls). Summary-level data for NAFLD were available from a GWAS meta-analysis (8434 cases and 770,180 controls), and data for 12 NAFLD-related features from corresponding published GWASs. The inverse variance weighted (IVW) analysis was employed as the primary statistical method. Bidirectional MR and CAUSE analysis were conducted to avoid reverse causality and false positive findings. In univariable MR analyses, we found evidence to support a causal effect of genetically predicted sleep apnea on NAFLD (OR = 1.50, 95% CI = 1.18-1.91) and HDL-C (β = -0.045, 95% CI = -0.090 to -0.001). In reverse MR, genetically predicted serum TG was associated with an increased risk of sleep apnea (OR = 1.07, 95% CI = 1.02-1.12), while genetically predicted HDL-C was associated with a decreased risk of sleep apnea (OR = 0.93, 95% CI = 0.89-0.98). After adjusting body mass index or educational attainment, none of these causal associations were retained. However, CAUSE method and MR analyses focusing on lipoprotein subfractions supported a causal effect of sleep apnea on HDL-C and HDL subfractions. This MR study indicated that sleep apnea has no direct causal association with NAFLD, elevated liver enzymes and insulin resistance. Our results showed suggestive inverse associations of genetically predicted sleep apnea on HDL-C and HDL subfractions, indicating that both HDL-C levels and HDL function may be causally implicated in sleep apnea.