Objective: To evaluate the benefits and harms of pediatric Tui Na as a non-pharmaceutical Chinese medicine therapy for acute diarrhea in children under 5 years of age. Design: Systematic review and meta-analysis of randomized clinical trials. Methods: We searched seven major English and Chinese databases from their inception to January 2018 for randomized clinical trials (RCTs) comparing pediatric Tui Na therapy with conventional medicine (montmorillonite/diosmectite or probiotics used alone or in combination). Two authors extracted data and assessed the Cochrane risk of bias, independently. The primary outcomes are clinical cure rate and diarrhea duration from admission to the cessation of diarrhea. 'Clinical cure' is defined as the frequency, timing and character of stool back to normal status, as well as disappearance of diarrhea symptoms. We present dichotomous data as risk ratio (RR), and continuous data as mean difference (MD) with their 95% confidence interval (CI). We used the Cochrane's Revman software (v.5.3) for data analysis. Trial sequential analysis (TSA) was applied to calculate the required sample size in a meta-analysis and detect the robustness of the results. The GRADEpro was used to generate a summary of finding table. Results: Totally 26 RCTs were included, involving 2410 children with acute diarrhea. Most of the included trials had high or unclear risk of bias in terms of random sequence generation, blinding, and incomplete outcome reporting. The pooled results demonstrated that pediatric Tui Na was superior to montmorillonite after three session treatment (RR 1.45, 95% CI 1.29-1.62, n = 772, 10 trials), and also superior to montmorillonite combined with probiotics after three-session treatment (RR 2.04, 95% CI 1.49-2.78, n = 533, 7 trials) and after six session treatment (RR 1.52, 95% CI 1.34-1.73, n = 631, 5 trials) in improving clinical cure rate. Pediatric Tui Na significantly decreased the duration of acute diarrhea (hrs) (MD -0.40 h, 95% CI -15.31 to -5.48 h, n = 410, 6 trials) and daily stool frequency (MD -1.71times, 95% CI -2.37 to -1.04, n = 217, 3 trials, after three session treatment). No adverse event related to pediatric Tui Na was reported in the included trials. The quality of evidence of included trials was generally moderate to low. TSA for cure rate demonstrated that the pooled data reached a sufficient power regarding both numbers of trials and participants. Conclusions: This review shows pediatric Tui Na appears to be effective and safe in improving clinical cure rate and shortening diarrhea duration in childhood aged less than five years of age with acute diarrhea. However, rigorously designed well-reported RCTs are warranted to confirm the findings.