The purpose of this study was to examine the reliability of the Japanese version of the Gross Motor Function Classification System (GMFCS) and to determine expert opinions on clinical use of this system using a Delphi survey. The reliability study was performed with 334 children (191 boys, 143 girls) with cerebral palsy, ranging in age from 8 months to 12 years (mean, 5 years 7 months ; standard deviation, 3 years 1 month). A total of 181 assessors participated in the study. Two assessors classified each child's level of gross motor function independently using a revised version of the GMFCS (Japanese version 1.1). This revision of the GMFCS was based on the results of previous pilot studies performed in Japan. A questionnaire was used for the Delphi survey, and the rate of positive response was calculated from the answers of 20 assessors at each institute that conducted the reliability study. In the reliability study, overall kappa was 0.67, but specific kappas <0.40 were found at level III and IV in the 4.6 year age group. In the Delphi survey, the rate of positive responses was not ≥80% only for the description of level III among the five levels. These findings and structural analysis of descriptions for level III and IV according to the results reported by Rosenbaum and coworkers suggest that reliability of the GMFCS was partly lowered because of the level III description for the age of 4.6 years, which might be set at a relatively lower level than actual development.