Objective This study aims to develop a decision support system for the non-drug management of gestational diabetes mellitus(GDM),conduct a preliminary application study,evaluate its usability and accuracy in clinical practice,and lay the foundation for a randomized controlled trial to explore its application effects. Methods A research group on GDM was established to construct a decision support system for non-drug management of GDM. The man-machine interaction technology was employed to develop 5 modules,including patients’ personal information,monitoring and feedback of blood glucose,diet,weight,and exercise records.The convenience sampling method was used to select GDM patients in the obstetrics nutrition clinic of a obstetrics and gynecology hospital in Shanghai in March 2022 as the research objects, and the systematic preliminary application was carried out. The data on using the system were analyzed,and surveys were conducted to evaluate the usability and structured evaluation of the system among medical staff,research groups,and patients. Results After initial application,the system was 100% effective in identifying abnormal blood glucose levels and sending immediate treatment tips,and 51% effective in generating blood glucose warning and prevention reports. The usability evaluation scores of GDM patients for the system and the structured evaluation scores of the system construction team members and health care workers were 88.15 and 90.63,respectively,which were both at a high level. Conclusion The system can provide accurate and objective decisions for medical staff and patients,and the preliminary application results reveal high accuracy and good usability. Additional studies can be conducted to investigate its effect on improving patients’ blood glucose levels and pregnancy outcomes to enhance the existing non-drug management model for GDM.