Objective: To examine (i) the prevalence of and associations between breastfeeding initiation and continuation by maternal diabetes status and (ii) the reasons for not initiating and/or continuing breast-feeding by maternal diabetes status. Design: Secondary data analyses of a population-based cross-sectional study were conducting using data from the US Centers for Disease Control and Prevention's Pregnancy Risk Assessment Monitoring System (PRAMS), 2009-2011. Multivariable logistic regression was used to investigate the associations between breast-feeding initiation and continuation by diabetes status. Setting: Thirty states and New York City, USA. Subjects: Mothers of recently live-born infants, selected by birth certificate sampling. Results: Among 72 755 women, 8·8 % had gestational diabetes mellitus (GDM) and 1·7 % had pregestational diabetes mellitus (PDM). Breast-feeding initiation was similar among GDM and no diabetes mellitus (NDM) women (80·8 % v. 82·2 %, respectively, P = 0·2), but continuation was lower among GDM (65·7 % v. 68·8 %, respectively, P = 0·01). PDM women had lower initiation and continuation compared with NDM (78·2 %, P = 0·03 and 60·4 %, P < 0·01, respectively). In adjusted analyses, current smoking status was a significant effect modifier for initiation, but not for continuation. Conclusions: Differences in breast-feeding initiation and continuation prevalence by maternal diabetes status may reflect differences in prenatal education, indicating the need for increased efforts among PDM women. Additionally, nonsmoking women with PDM or GDM would benefit from additional breast-feeding education.