The objective of this study was to analyse representative dietary patterns during pregnancy in Shanghai and explore the effects of dietary patterns during pregnancy on preterm birth. Data were derived from the ‘Iodine Status in Pregnancy and Offspring Health Cohort’ (ISPOHC) study. Multistage, stratified random sampling was used to select survey participants from 16 districts in Shanghai, which were divided into five sampling areas; 40–70 pregnant women were selected from each area. A total of 4361 pregnant women and their offspring were involved in the study. The male-to-female ratio of the babies was 1.04:1, and the incidence of single preterm birth was 4.2%. Three dietary patterns were extracted by factor analysis: a ‘Vegetarian Pattern’, an ‘Animal Food Pattern’ (AFP), and a ‘Dairy and Egg Pattern’. These patterns explained 40.513% of the variance in dietary intake. Binary logistic regression, which was used to analyse the association between birth outcomes and scores measuring maternal dietary patterns, found only the AFP was a significant risk factor for preterm birth. Higher AFP scores were positively associated with preterm birth (Q2 vs. Q1 OR = 1.487, 95% CI: 1.002–2.207; Q3 vs. Q1 OR = 1.885, 95% CI: 1.291–2.754). After adjusting for other potential contributors, a higher AFP score was still a significant risk factor for preterm birth (Q2 vs. Q1 OR = 1.470, 95% CI: 0.990–2.183; Q3 vs. Q1 OR = 1.899, 95% CI: 1.299–2.776). The incidence of preterm birth was 4.2%. A higher score of AFP was significantly associated with a higher risk of preterm birth. The animal food intake of pregnant women should be reasonably consumed during pregnancy.