Objectives To investigate predictors of preterm birth (PTB) in pregnancies with a short cervix and to create prediction models. Design Retrospective cohort study. Method Logistic regression model was used to identify predictors of PTB. The predictors were used to establish nomogram, which were validated using receiver operating characteristic (ROC) curve and calibration curve. Main outcome measures Preterm birth. Results Overweight or obesity (OR:2.00, 95% CI:1.114-3.51; OR:2.59, 95% CI:1.20-5.60), frequency of pregnancy ≧ 3 times (OR:1.97, 95% CI:1.14-3.40), twin pregnancy (OR:4.52, 95% CI:2.40-8.51), in vitro fertilization and embryo transfer (IVF-ET) (OR:2.24, 95% CI:1.19-4.19), gestational age at first diagnosis of short cervix (1st short cervix) (OR:0.953, 95% CI:0.910-0.999), cervical length (CL) at diagnosis of 1st short cervix (OR:0.908, 95% CI:0.86-0.96), history of PTB (OR:7.77, 95% CI:2.47-24.41), and autoimmune disease (OR:10.70, 95% CI:1.87-61.26) were predictors of PTB < 37 weeks, while twin pregnancy, gestational age of 1st short cervix, CL of 1st short cervix, history of PTB, and prepregnancy hypertension were predictors of PTB < 34 weeks. The area under the ROC curve (AUC) of the nomogram predicting PTB < 37 weeks and PTB < 34 weeks were 0.803 and 0.771, respectively. Both models showed good discrimination. Conclusions Gestational age of 1st short cervix, CL of 1st short cervix and other factors are strong predictors of PTB in pregnancies with a short cervix. Nomogram showed good discrimination and calibration, and hence might be effective in predicting risk of PTB for pregnancies with a short cervix. Keywords nomogram; preterm birth; risk factors; short cervix