The General Well-Being Schedule (GWBS) is a widely used, self-administered questionnaire to measure subjective well-being. It contains 18 items and was originally hypothesized to have six domains; however, the following studies suggest other factorial models. We devised the schedule's Japanese edition following the standard procedure and tested its psychometric properties considering Japanese cultural characteristics. Subjects were 1224 middle-aged participants undergoing health checkups in Adachi, Tokyo. The reliability of the original version was high: alpha coefficient, 0.90; test-retest Pearson's correlation coefficient, 0.81; and intra-class correlation coefficient, 0.85. Confirmatory factor analyses supported the structural validity of the 6-factor, the following 3-factor and 4-factor models (goodness of fit index: 0.932, 0.946 and 0.937, respectively). A new 3-factor model developed after the exclusion of item 11, which had no correlation with the other variables, was proved to have a higher fitness index (0.958) than the other models. This 17-item version showed adequate reliability (α coefficient, 0.91; test-retest Pearson's correlation coefficient, 0.81; and intra-class correlation coefficient, 0.85) and concurrent validity compared with the General Health Questionnaire, State-Trait Anxiety Inventory, Center for Epidemiologic Studies Depression Scale, Self-rating Depression Scale and Profile Of Mood State (correlation coefficients: -0.75 to -0.53). The median time required to complete the schedule was five minutes. Completing this instrument was easy for 80% of the subjects. These findings attest to the adequacy of the 17-item Japanese GWBS for the comprehensive assessment of individual inner state.