This study aims to explore the link between dietary habits and adverse pregnancy outcomes (APOs), including preterm birth (PB), preeclampsia (PE), gestational diabetes mellitus (GDM), fetal growth restriction (FGR), and spontaneous abortion (SA) through two-sample Mendelian randomization (MR). We accessed publicly available genome-wide association studies' (GWAS) summary statistics for dietary habits and APOs, respectively. We used five MR methods to synthesize MR estimates across genetic instruments. To ensure the robustness of our results, we assessed heterogeneity, and horizontal pleiotropy, and conducted sensitivity analyses. The primary analysis showed that intake of dried fruit (odds ratio (OR), 0.522; 95% confidence interval (CI): 0.291-0.935) and fresh fruit (OR, 0.487; 95% CI: 0.247-0.960) was related to a decreased risk of PB. While intake of tea (OR, 1.602; 95% CI: 1.069-2.403) and poultry (OR, 6.314; 95% CI: 1.266-31.488) was linked to a heightened risk of PB. Cheese intake was a protective factor against PE (OR, 0.557; 95% CI: 0.337-0.920) and GDM (OR, 0.391; 95% CI: 0.270-0.565). Intake of lamb/mutton had a negative relationship with PE (OR, 0.372; 95%CI: 0.145-0.954), whereas oily fish consumption showed a positive relationship with FGR (OR, 2.005; 95% CI: 1.205-3.339). However, after correction using the false discovery rate (FDR) analysis, only the intake of cheese showed a significant causal relationship with GDM (p < .001). Our study preliminarily found that cheese intake was significantly associated with the lower risk of GDM, while others were suggestively associated with the risk of APOs. Well-designed prospective studies are still needed to confirm our findings in the future.