Objective: To evaluate the application of three-in-one intelligent screening in outpatient pre-inspection in children's hospital. Methods: We randomly enrolled 100 children pre-screened by traditional method in the outpatient department of Children's Hospital of Zhejiang University from February 6th to 16th, 2020, and another 100 children by the intelligent three-in-one mode from February 17th to 27th, 2020. The traditional triage was conducted by nurses based on face-to-face, one-by-one interview of the epidemiological history and consultation department, and the temperature was measured before manual triage. The intelligent three-in-one model combined online rapid pre-inspection and triage, on-site manual confirmation, as well as synchronized online health education system. For on-line registered patients, the system automatically sent the COVID-19 epidemiological pre-screening triage questionnaire one hour before the appointment, requiring parents to complete and submit online before arriving at the hospital. The on-site registered patients were controlled at 100 m away from the hospital entrance. The nurses guided the parents to scan the QR code and fill in the COVID-19 epidemiological pre-examination triage questionnaire. At the entrance of the hospital, the nurse checked the guidance sheet and took the temperature again. The children with red guidance sheet were checked again and confirmed by pre-examination nurses, and accompanied to the isolation clinic through COVID-19 patients-only entrance. The children with yellow guidance sheet were guided to fever clinic. The children with green guidance sheet could go with their parents to the designated area, and then went to the corresponding consultation area. Health education was carried out throughout the treatment, and the system automatically posted the corresponding outpatient instructions and education courses. Parents would read the courses on their mobile phones and counsel online. The time of pre-examination and the coincidence rate of triage were compared between the two groups. Results: The three-in-one intelligent pre-inspection mode took an average of (25.6±8.0) s for each child, which was significantly shorter than the traditional pre-inspection mode (74.8±36.4) s (t=13.182, P < 0.01). The triage coincidence rate of the intelligent pre-inspection model was 98%, which was similar to that of traditional model (97%, χ²=0.251, P>0.05). Conclusion: The three-in-one intelligent pre-inspection model can effectively shorten the patient pre-check time, with similar triage coincidence rate to traditional model.