Recent studies have suggested a potential link between COVID-19 and the initiation or exacerbation of membranous nephropathy (MN) caused by the SARS-CoV-2 virus. However, a causal relationship between COVID-19 and MN remains unconfirmed. Here, we used the Mendelian randomization (MR) approach to investigate whether this causal relationship exists. We utilized data from the COVID-19 Host Genetics Initiative, comprising the largest available genome-wide association study (GWAS) datasets from European and East Asian populations. For MN, we used the largest literature-reported GWAS dataset. We selected single nucleotide polymorphisms as instrumental variables and used the inverse variance weighted (IVW) method, supplemented with MR Egger, weighted median, and mode methods. The MR-Egger intercept test and the MR-PRESSO global test were applied to assess horizontal pleiotropy. The robustness of the MR findings was evaluated using Cochran’s Q test and leave-one-out analysis. In both cohorts, MR methods showed SARS-CoV-2 infection, hospitalization, and severe COVID-19 did not increase the risk of MN. Conversely, MN potentially increased the risk of COVID-19-related hospitalization in Europeans, supported by the IVW and other methods. In East Asians, MR Egger suggested an increased risk of severe COVID-19 associated with MN, but other MR methods did not support this. No significant evidence of horizontal pleiotropy was detected. Cochran’s Q test showed no heterogeneity, and leave-one-out analysis confirmed the MR findings' reliability. The bidirectional MR analysis confirmed that COVID-19 did not increase MN risk, but MN might increase the risk of COVID-19-related hospitalization in Europeans, indicating a potential causal relationship.