The Child Health Record (CHR) is an essential document for children's health care activities. A cross-sectional study was performed with simple random sampling, aimed at analyzing factors related to quality of data completion on the CHR. Interviews and direct verification of 365 CHRs from 9 health districts in Belo Horizonte, Minas Gerais State, Brazil, were performed. A scoring system was created to classify the CHRs in terms of quality of data completion and to relate them to explanatory variables. Odds ratios were calculated by logistic regression. Completion of the 20 scoring items varied from 3.1% (use of iron supplements) to 99.7% (date of birth). Factors associated with worse scores were: children > 12 months old (OR = 1.77), mothers with < or = 6 years of schooling (OR = 1.97), children not treated by general practitioners (OR = 3.18), and mothers who had not received explanations on the CHR while in the maternity ward (OR = 1.77). The results show poor use of the CHR and emphasize the need for on-going efforts to train health professionals and prepare health services for their important role in the promotion of children's health.