Background: Observational studies have shown that high levels of serum uric acid (UA) were associated with atrial fibrillation (AF). However, the causal effect of urate on the risk of AF is still unknown. To clarify the potential causal association between UA and AF, we performed a Mendelian randomization (MR) analysis using genetic instrumental variables (IVs). Materials and methods: From the Korean GWAS dataset of 633 patients with AF (mean age 50.6 ± 7.8 years, 80.9% male, Yonsei AF Ablation cohort) who underwent radiofrequency catheter ablation and the data from 3533 controls (from the Korea Genome Epidemiology Study), we selected 9 SNPs, with a P value less than .05, associated with an increased UA serum level. Additionally, we calculated the weighted genetic risk score (wGRS) using the selected 9 SNPs, to use it as an instrumental variable. A Mendelian randomization analysis was calculated by a 2-stage estimator method. Results: The conventional association between the serum UA and AF was significant (P = .001) after adjusting for potential confounding factors. The SNP rs1165196 on SLC17A1 (F-statistics = 208.34, 0.18 mg/mL per allele change, P < .001) and wGRS (F-statistics = 222.26, 0.20 mg/mL per 1SD change, P < .001) were significantly associated with an increase in the UA level. The MR analysis was causally associated with rs1165196 (estimated odds ratio (OR), 0.21, 95% confidence interval (CI), 0.06-0.75, P = .017), but not wGRS (estimated OR, 1.07, 95% CI, 0.57-2.01, P = .832). Conclusion: The serum UA level was independently associated with the AF risk.