OBJECTIVE: The goal of this study was to evaluate tobacco-related documentation in children's medical records. METHOD: A cross-sectional, consecutive sample of 4216 parents at 13 primary care practices was surveyed on demographics, health habits, and smoking status of household members. The medical records of 2085 children from a subsample of 1149 families (all households with smokers and a sample of nonsmoking households) were reviewed for tobacco-related documentation at the first visit to the practice and visits in the 14 months preceding recruitment. Relationships of documentations with visit type, household smoking status, and use of charting prompts were examined. RESULTS: Most children (93%) had > or =1 visit during the reviewed period (77% had a health supervision visit), 23% were aged > or =11 years, 52% were Medicaid/uninsured, and 70% lived with smokers; 30.6% of children had family tobacco use status documented at a first visit to the practice and 15.4% had prenatal tobacco use status documented. Among children with a visit in the reviewed period, 39.3% with a health supervision visit and 9.6% without had a tobacco-related notation at a visit (P < .001). Overall, 15.2% of children living with a smoker had a visit notation indicating that someone in the household smoked. In households with smokers, documentation of household tobacco use status often disagreed with parent survey. Charting prompts significantly increased rates of identification of family tobacco use history and prenatal tobacco use history. CONCLUSIONS: Correct identification of household smoking status was absent for most children living with smokers. Improved documentation systems may facilitate tobacco-related surveillance and counseling.