Climate scenario dataset are indispensable for assessing future climate impacts. In this study, we developed statistically downscaled climate scenarios in Japan using modified bias correction method based on five general circulation models selected from the Coupled Model Intercomparison Project (CMIP) Phase 6 to facilitate impact assessments and adaptation strategies. Modification of time window of the original correction method results in successful agreement with the observed seasonal change of variables in each grid. The original CMIP6 models have a relatively small bias compared to CMIP5 models. The CMIP6-based bias-corrected scenarios are available for use with the emissions scenario of representative concentration pathway (RCP) 4.5 in addition to RCP2.6 and RCP8.5. Several temperature-related indices derived from the CMIP6-based climate scenarios agreed well with observations. The number of extremely hot days and nights increased nonlinearly in the future with additional global warming. An increase in the global warming level from 1°C to 2°C above the early 1900s would increase the probability of the number of extremely hot days per year exceeding the 2018 case by 4.1 times. The development of bias-corrected climate scenarios facilitates the study of various climate impacts on a CMIP6 basis.