This study was the first to evaluate the validity and reliability of molecular family stability as measured by a self-report instrument originally created for the U.S. population (Israel & Roderick, 2001), the Stability of Activities in the Family Environment (SAFE) measure, in Japanese children and caregivers. The Japanese version of the SAFE (J-SAFE) was previously developed through back-translation, committee evaluations, and bilingual testing (Maruyama, 2003). Participants for the present study were drawn from a cohort of elementary school children, age 8 to 12 years, from Toyama and Chiba, Japan. The sample included 291 caregivers and 205 children. Key findings indicated the J-SAFE is a reliable measure, with preliminary support for its validity. Reliability was high for both child and caregiver versions of the J-SAFE as demonstrated by Cronbach's alpha coefficients of .85 and .81, respectively, and by one-week test-retest reliability coefficients of .76 and .73, respectively. Results of confirmatory factor analyses provided support for a two-factor model of the J-SAFE for both the caregiver and child versions of the measure. Concurrent validity was demonstrated by the significant correlation between the caregiver J-SAFE and the CBCL Internalizing and Externalizing problem scores. In contrast, concurrent validity of the child J-SAFE was not supported through the correlation with the CBCL scores. Interestingly, caregivers of boys reported higher regularity in the family environment than caregivers of girls, whereas boys themselves reported lower regularity in the family environment than girls. The results from this study contribute to developing a better understanding of the measurement and relevance of molecular family stability for Japanese children and families.