Objective To compare the ability of six scoring systems to predict successful labor induction with cervical ripening among women with a Bishop score <3. Methods Secondary analysis of data from a prospective, multicenter observational Methods of Induction of Labor and Perinatal Outcomes (MEDIP) cohort study in 94 obstetrical French units. We included women with a Bishop score <3 before cervical ripening. We compared six scores: Bishop, simplified Bishop, modified Bishop, simplified Bishop including parity, Hughey, and Levine scores. Vaginal delivery defined successful labor induction. The ability of each score to predict successful labor induction was evaluated by comparing their area under the curve (AUC). Results Among the 600 eligible women in this study, 408 (68%) delivered vaginally. Body mass index (calculated as weight in kilograms divided by the square of height in meters; mean ± standard deviation [SD]: 24.7 ± 5.5 vs 26.0 ± 5.7; P = 0.01) and nulliparity (48.8% vs 85.4%; P < 0.001) were lower in the successful induction group, whereas height was higher (mean ± SD: 165.3 ± 6.0 cm vs 163.7 ± 6.0 cm; P = 0.002). The simplified Bishop including parity, Hughey, and Levine scores had the highest AUC (0.70, 95% confidence interval [CI] 0.65–0.73; 0.68, 95% CI 0.64–0.74; and 0.69, 95% CI 0.65–0.74, respectively). Conclusion In women with a very unfavorable cervix, scores that include parity predict successful labor induction more accurately, such as simplified Bishop including parity, Hughey, or Levine scores.