Abstract: Objective   To establish a system to prevent research misconduct from the perspective of editorial office of medical sci-tech journals, so as to reduce the phenomenon of research misconduct in medical disciplines and improve the journal reputation. Methods  Editors, faculty, administrators, and students were deeply interviewed, and questionnaires were also conducted. SPSS 22 software was used to process the data and t test was adopted for comparison. Cronbach's α coefficient of this survey is 0.736, indicating a high internal consistency. It means the questionnaire is reliable. The preliminary framework was verified and improved based on the survey result of the cognition and current situation of research misconduct education and prevention. Results  Most of the respondents (> 91%) had a clear understanding of academic misconduct, and most of the respondents (> 85%) had a correct attitude towards academic misconduct. Among the research misconducts frequently encountered by editors, the most popular choice appeared to be “no ethical approval or unproven” for editors who had worked for 11 to 15 years (P=0.039), and the difference was statistically significant. The junior and intermediate editors were more prone to the establishment of the system of changing author signature, order and retraction(P=0.025). The proportion of editors with junior and intermediate titles who chose publishing announcement in the punishment measures for research misconduct was significantly higher than that of editors with senior titles (P=0.001), and the difference was statistically significant. Conclusion  There are many causes for research misconduct. Therefore, it is particularly important not only to strengthen the moral self-discipline of the authors, but also to strengthen the supervision of the society, improve the relevant legal system and build a scientific and perfect research misconduct prevention system. This paper discussed the prevention system of research misconduct in the editorial office of medical journals from five levels: moral education, legal basis, system construction, technical control and recognition and punishment, which is conducive to preventing the production of research misconduct papers.

Key words: research misconduct, prevention system, medical science journals, editorial office