Purpose <p>Our aim is to evaluate the causal association between medium-chain acyl-CoA dehydrogenase (MCAD) and the incidence of diabetic kidney disease (DKD) through bidirectional Mendelian randomization (MR) analysis.</p> Methods <p>We obtained MCAD and DKD data from the open-source genome-wide association studies (GWAS) databases, the GWAS Catalog and the FinnGen consortium, and selected instrumental variables (IVs). Bidirectional MR analyses were conducted to assess the causal association between MCAD and DKD, using the inverse variance weighted (IVW) method as the primary approach. F-statistics were applied to evaluate the quality of IVs. Sensitivity analyses were used to ensure the robustness of the results.</p> Results <p>Forward MR analysis suggests that MCAD is a protective factor against DKD in Europeans. The average F-statistic was 213.90, indicating that the selected single-nucleotide polymorphisms (SNPs) were strong IVs. There was no evidence of horizontal pleiotropy (<i>P</i> = 0.524) or heterogeneity (Q-pval = 0.894) among the IVs. The reverse MR analysis did not support a reverse causal association between DKD and MCAD (OR = 1.0281, 95% CI 0.9896–1.0680, <i>p</i> = 0.155). The IVs used in this process had an average F-statistic of 57.94, confirming them as strong IVs. Additionally, the results were robust, showing no evidence of horizontal pleiotropy (<i>P</i> = 0.914) or heterogeneity (Q-pval = 0.658).</p> Conclusions <p>There is a causal association between MCAD and the incidence of DKD, suggesting that MCAD expression may serve as an important protective factor against DKD. Conversely, no significant causal effect of DKD on MCAD expression was observed.</p>