Objectives: Previous observational studies have revealed a connection between sarcopenia and COVID-19. To evaluate their causal relationship, we utilized a bidirectional two-sample Mendelian randomization (MR) analysis to study the link of cause and effect between sarcopenia and COVID-19. Methods: Inverse variance weighting (IVW), MR-Egger, weighted, and weighted median were used in this research. Then we used the MR Pleiotropy RESidual Sum and Outlier (MR-PRESSO) and MR-Egger regression methods to estimate the pleiotropy of instrumental variables (IVs), while the outliers were excluded by MR-PRESSO. Moreover, we used Cochran’s Q statistic to evaluate the heterogeneity among the IVs. And we used leave-one-out sensitivity analysis to identify the SNPs that significantly affect the outcomes. Finally, the Bonferroni correction was used to correct each result. Results: The IVW results suggested that faster WP decreased the risk of all types of COVID-19 (COVID-19 infection: OR = 0.469, 95% CI: 0.326,0.676, P = 4.82E-05; COVID-19 hospitalization: OR = 0.247, 95% CI: 0.122,0.502, P = 1.11E-04; severe COVID-19: OR = 0.120, 95% CI: 0.046,0.314, P = 1.53E-05). However, there was no causal relationship between ASM, LH or RH and COVID-19, and WP adjusted for BMI had no significant connection with all types of COVID-19. Furthermore, no causal association of COVID-19 on sarcopenia was observed in the results of reverse MR analysis. Conclusion: Our bidirectional two-sample MR study suggests the causal relationship between WP and COVID-19 but it may be caused by the mediating role of BMI, thus there is no causal association between sarcopenia and COVID-19.