Previous evidence suggests that dietary intake can affect liver diseases; However, the causal relationship between dietary intake and liver diseases remains unclear. To investigate this, we conducted a bidirectional Mendelian randomization (MR) analysis to comprehensively assess the potential causal relationship between dietary intake and liver diseases. Two-sample bidirectional MR was performed based on genome-wide association studies summary data from the UK Biobank and FinnGen database. The primary analysis method for evaluating causal relationships was inverse-variance weighted. Supplementary analyses included MR-Egger and weighted median methods. Subsequently, sensitivity analyses were performed using Cochran Q test, MR-Egger intercept test, MR-PRESSO, RadialMR, and leave-one-out analysis to assess heterogeneity and horizontal pleiotropy. MR evidence indicated that genetically predicted poultry intake (adjusted odds ratio [OR] = 0.04, 95% confidence interval [CI] = 0.00–0.43, P = .007) and salad/raw vegetable intake (adjusted OR = 0.18, 95% CI = 0.04–0.83, P = .028) were directly associated with a reduced risk of cirrhosis. Conversely, there is no causal association between dietary intake and nonalcoholic fatty liver disease, alcoholic liver disease, or hepatocellular carcinoma. This study provides evidence supporting the impact of dietary intake on liver disease. Increased intake of poultry and salad/raw vegetables is associated with a reduced risk of cirrhosis. These findings can inform preventive and therapeutic strategies for cirrhosis.