Treatment complexity of cerebral palsy (CP) patients imposes outcome evaluation studies, which may include objective technical analysis and more subjective functional evaluation. The Edinburgh Gait Score (EGS) was proposed as an additive or alternative when complex instrumented three-dimensional gait analysis is not available. Our purposes were to apply a translated EGS to standard video recordings of independent walking spastic diplegic CP patients, to evaluate its intraobserver and interobserver reliability with respect to gait analysis familiar and not familiar observers.Ten standard video recordings acquired during routine clinical gait analysis were examined by eight observers gait analysis interpretation experienced or not, out of various specialities, two times with a two weeks interval. Kappa statistics and intraclass correlation coefficient were calculated.Better reliability was observed for foot and knee scores than in proximal segments with significant differences between stance and swing phase. Significantly better results in gait analysis trained observers underlines the importance to either be used to clinical gait analysis interpretation, or to benefit of video analysis training before observational scoring.Visual evaluation may be used for outcome studies to explore clinical changes in CP patients over time and may be associated to other validated evaluation tools.