Background <p>It has been shown that low testosterone levels are associated with the development of osteoarthritis (OA). In our study, we aimed to investigate a bidirectional causal relationship between bioavailable testosterone levels and OA using Mendelian randomization (MR) analysis.</p> Methods <p>In our study, the datasets from publicly available genome-wide association study (GWAS) were adopted, including the OA-related dataset (ukb-b-14486) and the bioavailable testosterone levels-related dataset (ebi-a-GCST90012104). The UKB-B-14,486 dataset contains 462,933 samples in total, including 38,472 OA samples, 424,461 control samples, and 9,851,867 SNPs, all collected from the European population in 2018. Additionally, the EBI-A-GCST90012104 dataset includes 382,988 samples and 16,137,327 SNPs, which reflect data from the European population in 2020. In total, five methods were utilized, namely MR Egger, Weighted median, Inverse variance weighted (IVW), Simple mode, and Weighted mode. Among them, IVW was the main analytical method. Additionally, the sensitivity analysis was carried out through the heterogeneity test, the horizontal pleiotropy test, and the Leave-One-Out (LOO) method.</p> Results <p>The result of forward MR analysis demonstrated that bioavailable testosterone levels were considerably relevant to OA, and were a risk factor for OA (OR = 1.01, 95% CI: [1.00, 1.02], <i>P</i> = 0.02). However, through reverse MR analysis, we did not find a causal relationship between OA and bioavailable testosterone levels. Moreover, the results of the sensitivity analysis suggested that our results were reliable.</p> Conclusion <p>The results of our study supported a causal relationship between bioavailable testosterone levels and OA.</p>