Objective: To evaluate the evidence of using Action Observation Therapy in the rehabilitation of children with Cerebral Palsy. Study design: Systematic review with meta-analysis of Randomised Controlled Trials. Methods: For the purpose of identifying relevant studies, six databases were searched from inception until July 2020. The methodological quality was assessed by Physiotherapy Evidence Database scale. The outcomes were classified within the framework of the International Classification of Functioning. A pooled meta-analysis was performed on studies that demonstrated homogeneity. Results: Twelve randomised controlled trials with 307 participants were included with six of them were included in the meta-analysis. Non-significant difference between the groups was demonstrated by meta-analysis. Results of capacity assessed in post treatment and follow up evaluation were (0.06, –0.22 to 0.34, 95% (CI); P = 0.69 and (–0.35, –0.96 to 0.27, 95% (CI); P = 0.27); respectively. Actual performance in post-treatment and follow up were (0.10, –0.22 to 0.48, 95% (CI); P = 0.62) and (0.01, –0.40 to 0.41, 95% (CI); P = 0.97); respectively. Perceived performance evaluated using (ABILHAND-KIDS) were (0.30, –0.28 to 0.89, 95% (CI); P = 0.31) and (0.15, –0.43 to 0.73, 95% (CI); P = 0.61) for post treatment and follow up; respectively. Overall effect on activity domain was (0.08, –0.11 to 0.28, 95% (CI); P = 0.86) immediately and (0.04, –0.33 to 0.26, 95% (CI); P = 0.49) at follow-up; respectively. Conclusion: No evidence of benefit had been found to draw a firm conclusion regarding the effectiveness of action observation therapy in the rehabilitation of children with cerebral palsy due to limitations in methodological quality and variations between studies.