[Objective] To investigate the management and factors related to closed reduction failure of humeral supracondylar fracture in children. [Methods] From June 2019 to December 2020, a total of 375 children with supracondylar fracture of humerus were treated in our hospital. All patients underwent closed reduction (CR) firstly. If the CR proved successful, the patients had percutaneous cross Kirschner wire for internal fixation conducted (the CR group), while the unsuccessful patients underwent open reduction (OR) and cross Kirschner wire fixation (the OR group). The clinical and imaging data were compared between the two groups. In addition, univariate comparison and binary logistic regression analysis of preoperative data based whether CR successful were performed to search the factors related to CR failure. [Results] Among 375 patients, 348 patients had CR performed successfully, accounting for 92.80%, whereas 27 patients got CR failed to change OR, accounting for 7.20%. Finally, all patients got fracture reduction and cross Kirschner wire fixation successfully in both groups. The CR group proved significantly superior to the OR group in terms of operation time, intraoperative fluoroscopy times, hospital stay, incidence of early complications and VAS score 1 day postoperatively (P<0.05). All the patients were followed up for 14~46 months, with an average of 23.40（±4.59）月months. There were no statistically significant differences between the two groups in the time to remove the plaster splint, the time to resume full weight-bearing activities, elbow flexion-extension ROM and FIynn’s scales at the latest follow-up (P>0.05). Radiographically, satisfactory fracture reduction achieved in all children of both groups, with sound fracture healing, whereas no significant differences in carrying angle and Baumann angle between the two groups at the last follow-up (P<0.05). Regarding to univariate comparison, the CR group had significantly less BMI, fracture complexity, time elapsed between injury and surgery, as well as damage energy extent than the OR group (P<0.05). As results of logistic regression, fracture complexity type (OR=8.251, P=0.037), damage energy (OR=1.593, P=0.035), and duration from injury to operation (OR=1.400, P=0.026) were independent risk factors for CR failure of supracondylar humeral fractures in children. [Conclusion] The children who fail to initial closed reduction of humeral supracondylar fracture should be changed to open reduction in time, and still archive satisfactory results eventually. The fracture complexity, high energy injury and long duration from injury to operation are the main risk factors for the failure of initial closed reduction.