This study sought to validate the Japanese version of the Survey Work–Home Interaction–NijmeGen, the SWING, which assesses multi-dimensional work–family interaction by differentiating between the direction and quality of influence. We translated the SWING into the Japanese language, the SWING-J. A back-translation procedure confirmed that the translation was appropriate. A total of 2701 dual-earner parents with preschool children (1193 men and 1508 women) were surveyed. The complete questionnaire included the SWING-J, job and family domain variables, and well-being indicators. The reliability and factorial and convergent validity of the used measures were examined. As the results, four dimensions (i.e. work-to-family negative spillover, family-to-work negative spillover, work-to-family positive spillover and family-to-work positive spillover) were determined by an exploratory factor analysis. A series of confirmatory factor analyses suggested that the hypothesized four-factor model provided a reasonably good fit to the data. Convergent validity was generally supported by the expected correlations of work–family spillovers with the possible predictors and consequences. Cronbach's alpha coefficients of the four subscales of the SWING-J were satisfactory (0.75–0.86). The present study confirmed that the Japanese version of the SWING is an adequate tool to measure positive and negative spillover between working life and family life among Japanese workers.