OBJECTIVE: To assess the prevalence of and factors associated with actual recent practice and near-future intention for infant sleep location in a national sample. METHODS: There were 3260 mothers from 32 US hospitals who responded to a survey at infant age 2 to 6 months regarding care practices, including usual and all infant sleep locations in the previous 2 weeks and intended location for the next 2 weeks. Mothers were categorized as (1) having practiced and/or intending to practice exclusive room-sharing without bed-sharing, (2) having practiced anything other than exclusive room-sharing but intending to practice exclusive room-sharing, (3) intending to have the infant sleep in another room; and (4) intending to practice bed-sharing all night or part of the night. Multivariable multinomial logistic regression examined associations between sleep-location category, demographics, feeding method, doctor advice, and theory of planned behavior domains (attitudes, social norms, and perceived control). RESULTS: Fewer than half (45.4%) of the mothers practiced and also intended to practice roomsharing without bed-sharing, and 24.2% intended to practice some bed-sharing. Factors associated with intended bed-sharing included African American race and exclusive breastfeeding; however, the highest likelihood of bed-sharing intent was associated with perceived social norms favoring bed-sharing (adjusted odds ratio [aOR] 5.84; 95% confidence interval [CI] 4.14-8.22) and positive attitudes toward bed-sharing (aOR 190.1; 95% CI 62.4-579.0). Women with a doctor's advice to room-share without bed-sharing intended to practice bed-sharing less (aOR 0.56; 95% CI 0.36-0.85). CONCLUSIONS: Sleep-location practices do not always align with the recommendation to roomshare without bed-sharing, and intention does not always correspond with previous practice. Attitudes, perceived social norms, and doctor advice are factors that are amenable to change and should be considered in educational interventions.