Objective: To evaluate the predictive value for successful labor induction of transvaginal ultrasound (TVS) of cervical length according to parity. Method: TVS of the cervix was performed before term labor induction. Induction was considered successful if vaginal delivery was achieved within 24 hours; 231 women were available for final analysis. Results: Analysis of the receiver operator characteristics curve showed an optimal cut-off for cervical length of ≤20 mm for successful induction. Following multivariate logistic regression analysis, a sonographic short cervix (AOR 5.6; p < 0.001) was an independent predictor of successful induction but not a favorable Bishop score (p = 0.47). Among multiparas with a short cervix, positive and negative predictive values for successful induction were 98% (95% CI 90-100%) and 21% (95% CI 13%-32%) and among nulliparas, predictive values were 69% (95% CI 53%-82%) and 77% (95% CI 64%-87%) respectively. Conclusion: In nulliparas, cervical length can usefully predict labor induction outcome.