Background <p>Previous studies have indicated that retinol and vitamin D may be associated with the oncogenesis of tongue cancer. Therefore, we aimed to assess the causal relationships of retinol and vitamin D with the risk of tongue cancer using the two-sample Mendelian randomization (MR) method.</p> Methods <p>Single nucleotide polymorphisms (SNPs) related to retinol, vitamin D and tongue cancer were obtained from the up-to-date genome-wide association study (GWAS) catalogue, which was screened for instrumental variables (IVs). We performed two-sample MR analyses and used inverse-variance weighted (IVW) as the primary method. Additionally, we used the MR-pleiotropy residual sum and outlier (MR-PRESSO) method, MR-Egger intercept analysis, Cochran’s Q test and leave-one-out analysis to evaluate the sensitivity of MR.</p> Results <p>The IVW method revealed that retinol was not significantly correlated with the risk of tongue cancer (OR = 0.8602; 95% CI = 0.4453–1.6617; <i>P</i> = 0.654). However, the causal relationship between vitamin D and the risk of tongue cancer was significant according to IVW (OR = 0.4003; 95% CI = 0.1868–0.8577; <i>P</i> = 0.019). The sensitivity analysis did not detect any significant horizontal pleiotropy or heterogeneity.</p> Conclusions <p>Given the limitations of this study, our MR study suggests that retinol is unlikely to influence the risk of tongue cancer, but vitamin D may decrease the risk of tongue cancer.</p>