Although assessment of the quality of movement in children with cerebral palsy (CP) is difficult, the development of the Gross Motor Performance Measure (GMPM) has facilitated this process. In order to determine the interobserver reliability of the GMPM, 36 children with spastic neuromuscular disorders (mean age 7 years, range 4 to 15 years) were evaluated using four of the five dimensions of the GMPM. Percent Agreement, Intraclass Correlations, and Kappas were calculated by both dimension and attribute to determine reliability. In addition, reliability measures were evaluated over time to determine whether reliability improved with continual use of the GMPM. Overall, interobserver reliability was in the ‘fair to good’ category regardless of the reliability measure used in the analysis. Reliability scores improved over time with a greater number of individual item scores moving from the ‘fair to good’ category to the ‘excellent’ category. Results from this study indicate that it is possible to assess reliably the quality of movement in children with CP.