To examine reliability, validity, factor structure, and measurement invariance (i.e., configural, metric, and scalar invariance) of the Japanese Conners' Adult attention deficit hyperactivity disorder (ADHD) Rating Scales (CAARS), Japanese nonclinical adults (N = 786) completed the CAARS Self-Report (CAARS-S). Each participant was also rated by one observer using the CAARS Observer Form (CAARS-O). For the test of measurement invariance, excerpts from the original (North American) CAARS norming data (N = 500) were used. Dimensional structure was examined by exploratory and confirmatory factor analyses. Test-retest reliability, internal consistency, and concurrent validity were satisfactory. Based on the "DSM-IV" model and Japanese four-factor model, configural and metric invariance were established for the CAARS-S/O across Japanese and North American populations. Scalar invariance was established for the CAARS-O based on the Japanese model. The use of the Japanese CAARS for diagnostic purposes in Japan was supported; however, it should be used with caution for cross-cultural comparison research.