This study aimed to evaluate the psychometric properties of the 20-item Japanese version of the Dutch Eating Behavior Questionnaire for Children (DEBQ-C). A population-based sample of 502 Japanese school children who were fifth- and sixth-graders in elementary schools and first- and second-graders in junior high schools was assessed. A sample was randomly split into two subsamples, one for exploratory factor analysis (EFA) and another for confirmatory factor analysis (CFA). Internal consistency estimates for subscales (restrained, emotional, and external) were evaluated using Cronbach's alpha. Measurement invariance was examined across each subgroup (genders: boys and girls, school categories: elementary school and junior high school, body mass index (BMI) categories: underweight, normal weight, and overweight) by using multi-group CFA. The Japanese version of the DEBQ-C demonstrates good results of item analysis. The three-factor structure of the original DEBQ-C was supported by both EFA and CFA. The reliability of each factor was also satisfied (restrained: α = 0.86, emotional: α = 0.90, external: α = 0.86). Results of multi-group confirmatory factor analysis supported its metric and scalar or partial scalar measurement invariance across all subgroups. In gender subgroup, girls scored higher on restrained eating. In school subgroup, junior high school children scored higher on emotional and external eating. In BMI subgroup, overweight children scored higher on restrained eating. These findings suggest that the Japanese version of the DEBQ-C is a psychometrically valid and reliable instrument for assessing eating behaviors across gender, school categories, and BMI categories in Japanese children.