AIM The purpose of this study was to develop and validate the Japanese version of the Abbey Pain Scale (APS-J), to assess pain of older adults who live in nursing homes in Japan. METHODS In this study, the data were collected from residents in two nursing homes in Japan to include: demographics, the Barthel Index, Folstein Mini-Mental Examination (MMSE), APS-J and Verbal Descriptor Scale (VDS) for pain. Two researchers independently assessed the residents' pain using the APS-J while the residents walked or were transferred from bed to wheelchair. Intraclass correlation coefficients (ICC) for inter-rater and test-retest reliability, Chronbach's alpha-value of the APS-J, and correlation between the APS-J and other variables were examined. RESULTS Data were obtained from 171 residents. The ICC for inter-rater and test-retest reliability were 0.824 and 0.657, respectively. Internal consistency was 0.645 for the total sample and 0.719 for those with an MMSE score of 0 (n = 58). Multiple regression analysis showed that contractures (P < 0.001), previous injuries (P < 0.001), the MMSE (P = 0.003) and paralysis (P = 0.018) were independently associated with the APS-J. The APS-J and VDS were moderately correlated (r = 0.49; P < 0.01). The APS-J total score and subscales, "change in body language" and "physical changes", were significantly different among subjects with different MMSE score groups. CONCLUSION Findings show some evidence for the reliability and validity of APS-J.