Data on the relationship between the dose of opioid replacement therapy in pregnancy and the risk and severity of neonatal opioid withdrawal syndrome are conflicting and have methodological limitations. To assess the association of methadone and buprenorphine dose at delivery with neonatal opioid withdrawal syndrome in a large cohort. We performed a retrospective cohort study using data from a comprehensive perinatal opioid dependency program from 2000 through 2016. Women with a history of opioid use disorder enrolled in a medication-assisted treatment program were included. Strict neonatal opioid withdrawal syndrome case definition and neonatal treatment guidelines were utilized throughout the study epoch. Comparisons were made between women on methadone and buprenorphine. The dose of opioid replacement at delivery and the risk and severity of neonatal opioid withdrawal syndrome were assessed with univariable analysis and multivariable logistic regression. In all analyses, methadone and buprenorphine dosing were evaluated as a continuous variable. Four hundred eighty two of 709 women (68.0%) met inclusion criteria including 344 on methadone (71.4%) and 138 on buprenorphine (28.6%). Nonopioid polysubstance abuse, body mass index, medication-assisted treatment compliance, birthweight, and other characteristics were similar between groups. Overall, the frequency of neonatal opioid withdrawal syndrome was not significantly different between the methadone and buprenorphine groups (56.8% vs 52.0%, P= .35). Dose at delivery ranged at 0-165 mg for methadone and 0-30 mg for buprenorphine. In a univariable analysis, methadone dose at delivery was associated with neonatal opioid withdrawal syndrome (83.0 ± 34.2 mg vs 71.9 ± 35.8 mg for neonatal opioid withdrawal syndrome vs nonneonatal opioid withdrawal syndrome neonates, P < .001), but buprenorphine dose at delivery was not (8.4 ± 4.4 vs 7.6 ± 4.8 mg for neonatal opioid withdrawal syndrome vs nonneonatal opioid withdrawal syndrome neonates, P= .30). Peak neonatal opioid withdrawal syndrome score, duration of neonatal opioid withdrawal syndrome treatment, and cumulative neonatal morphine exposure were significantly associated with delivery methadone dose but not buprenorphine dose. The association between delivery methadone dose and neonatal opioid withdrawal syndrome persisted in multivariable regression. The dose of methadone at the time of delivery is associated with the frequency and severity of neonatal opioid withdrawal syndrome, with higher doses associated with more severe neonatal opioid withdrawal syndrome when analyzed continuously. These data may inform future prospective studies on methadone dosing in pregnancy. While medication-assisted treatment agent and dose may have an impact on pertinent neonatal outcomes related to neonatal opioid withdrawal syndrome, the provision of medication-assisted treatment in pregnancy should reflect the goal of prevention of recidivism and maternal mortality and utilize an approach that balances fetal and maternal risk to optimize outcomes.