The Parenting Stress Index-Short Form (PSI-SF) has been widely employed to assess parenting stress in a number of research and clinical trials. To date, no parenting stress studies in Japan have examined the factor structure, validity, and reliability of the PSI-SF. Therefore, this study aimed to evaluate the psychometric properties of this 19-item version as administered in a national cohort study, the Japan Environment and Children’s Study, to two sample groups of mothers, those with a 1.5-year-old child and those with a 2.5-year-old child (n = 79,282 and 75,831, respectively). We performed exploratory factor analysis to re-examine the appropriate factor structure, confirmatory factor analysis to evaluate goodness of fit, and calculated Cronbach’s α and Pearson’s r coefficients to evaluate internal consistency and reproducibility over time, respectively. The results highlighted that a three-factor structure fit the instrument better than a two-factor structure, yielding better scores for the model fit indices and the α and r coefficients. In addition, the third factor identified in this study was strongly associated with having a relationship with and help from the husband. The findings suggest the importance of using a parenting stress scale with various factors to evaluate mothers’ parenting stress.