Objective:To investigate the causal association between testosterone and nonalcoholic fatty liver disease(NAFLD) in men and women using a two-sample Mendelian randomization(MR) approach.Methods:Genetic variation in testosterone(total testosterone, bioavailable testosterone) and sex hormone-binding globulin(SHBG) in females and males was used as an instrumental variable using the genome-wide association study(GWAS) pooled data, and the inverse variance weighting method was applied. Inverse variance weighted(IVW) was used as the main analytical method, along with six univariate MR methods based on other modeling assumptions to assess the causal relationship between testosterone(total testosterone, bioavailable testosterone) as well as SHBG and NAFLD in women and men. In addition, NAFLD data from Finnish Biobank(FinnGen) were applied to validate the results of the exploratory analysis. Further, sensitivity analyses were performed to assess the level of heterogeneity, genetic pleiotropy, and stability of the instrumental variables using Cochran′ s Q test, MR-Egger regression, and leave-one-out methods. Results:The results of exploratory analysis of IVW model showed that bioavailable testosterone and SHBG were causally associated with NAFLD in women, for each unit increase in bioavailable testosterone levels, the risk of developing non-alcoholic fatty liver disease(NAFLD) rose by 24%( OR=1.24, 95% CI 1.07-1.43, P=0.004); and with each unit decrease in women′s SHBG, the NAFLD risk increased by 31%( OR=0.69, 95% CI 0.57-0.83, P<0.001). However, testosterone(total testosterone, bioavailable testosterone) as well as SHBG in men and female total testosterone did not show a causal relationship with NAFLD. The results of the other six MR methods were generally consistent with the IVW method. The results of the external validation data provided further evidence of a causal relationship between female bioavailable testosterone and SHBG and NAFLD. Conclusion:Elevated levels of bioavailable testosterone along lower levels of SHBG may increase the risk of developing NAFLD in women.