OBJECTIVE: To study the annual incidence of injuries and the relationship between behavior problems and injuries among school-age children. METHODS: A prospective cohort study on injuries for 1-year follow-up period was conducted among 2 005 school-age children selected by cluster sampling from three primary schools in Maanshan city. They subjects were divided into two groups with or without exposure according to behavior problems rated by the Rutter Child Behavior Questionnaire at the beginning of the study. Nonparametric test was performed to analyze the differences in injuries between the two groups of children, and the influential factors for injuries were analyzed with multi-classification ordinal response variable logistic regression model. RESULTS: The overall incidence rate for injuries in school-age children was 42.51%, while among children with and without behavior problems were 64.87% and 38.85%, respectively. There were significant differences between the two groups (u = -6.054, P = 0.000). However, the incidence rates of injuries in school-age children with antisocial (A) behavior, neurotic (N) behavior and mixed (M) behavior were 66.99%, 67.41% and 61.40%, respectively. No significant differences were found among them (u(A,N) = -0.052, P = 0.958; u(A,M) = -0.400, P = 0.689; u(N,M) = -0.364, P = 0.716). Multivariate analysis indicated that injuries in school-age children were associated with children behavior problems, maternal age at childbirth, bad conditions during mother pregnancy, education background of mother, prevention measures for safety at home and the child accompanied to travel between school and home by adults. Behavior problems of children seemed to be the major risk factors for injuries. CONCLUSION: Children with behavior problems represented a significant risk group for injuries among school-age children. When planning intervention strategies on injuries, behavior problems should be emphasized to ensure optimal effectiveness of intervention.