Background: Observational studies have found an association between increased whole body water mass (BWM) and atrial fibrillation (AF). However, the causality has yet to be confirmed. To provide feasible protective measures on disease development, we performed Mendelian randomization (MR) design to estimate the potential causal relationship between increased BWM and AF. Methods: We implemented a two-sample MR study to assess whether increased BWM causally influences AF incidence. For exposure, 61 well-powered genetic instruments extracted from UK Biobank (N = 331,315) were used as the proxies of BWM. Summary genetic data of AF were obtained from FinnGen (Ncase = 22,068; Ncontrol = 116,926). Inverse-variance weighted (IVW), MR-Egger and weighted median methods were selected to infer causality, complemented with a series of sensitivity analyses. MR-Pleiotropy Residual Sum and Outlier (MR-PRESSO) and Radial MR were employed to identify outliers. Furthermore, risk factor analyses were performed to investigate the potential mechanisms between increased BWM and AF. Results: Genetic predisposition to increased BWM was demonstrated to be significantly associated with AF in the IVW model (OR = 2.23; 95% CI = 1.47–3.09; p = 1.60 × 10−7), and the result was consistent in other MR approaches. There was no heterogeneity or pleiotropy detected in sensitivity analysis. MR-PRESSO identified no outliers with potential pleiotropy after excluding outliers by Radial MR. Furthermore, our risk factor analyses supported a positive causal effect of genetic predicted increased BWM on edematous diseases. Conclusions: MR estimates showed that a higher BWM could increase the risk of AF. Pathological edema is an important intermediate link mediating this causal relationship.