The aim of the study was to investigate the possible existence of a bifurcation point in sleep duration, which could differentiate children with behavioral problems from those without. We conducted a survey of children's sleep and health between June and July 2012. A total of 383 children aged 4-6 years were studied (193 boys, 190 girls; average age 4.7 years; 223 and 160 children attended kindergarten and childcare centers, respectively), based on their parents' reports using the Children's ChronoType Questionnaire and the Strength and Difficulties Questionnaire. The dependent variable was the dichotomized behavior of total behavioral problems; the cut-off value was >= 13 points. The independent variable was sleep duration. The bifurcation point of sleep duration was the median value: 9.5 h for sleep period, 10.0 h for time in bed (TIB), and 10.5 h for the sum of TIB and naptime (24 h TIB). Binomial multiple logistic regression analyses were performed separately for sleep period, TIB, and 24 h TIB, which were controlled for chronotype, sex, age, and childcare programs; those corresponded to Models 1, 2, and 3, respectively. The results show that there was no significant association between total behavioral problems and sleep duration. However, using morning-type as a reference, adjusted odds ratios (95 % confidence intervals) for neither-type and evening-type were 2.06 (1.12-3.76) and 3.72 (1.60-8.66) in Model 1, 2.09 (1.14-3.82) and 3.89 (1.71-8.88) in Model 2, and 1.89 (1.03-3.50) and 3.61 (1.55-8.40) in Model 3.