Objective: To address the poor prognosis of patients caused by the weak knowledge and non-standard operation of blood glucose management among non-endocrinology specialist nurses during the perioperative period, this study aims to establish a scientific and systematic structured training system. Through a training model that combines theory with practice, it aims to enhance the perioperative blood glucose management ability of nurses, improve the blood glucose control effect, and clinical outcomes of patients. Method: A mixed research method was adopted. The training theoretical framework was constructed through literature analysis. The weights of the course modules were determined by combining semi-structured interviews (n = 35) and two rounds of Delphi expert inquiries (expert authority coefficient 0.89) (first-level indicator weights: theoretical foundation 0.25, skill operation 0.35, clinical application 0.28, emergency treatment 0.12). Finally, a structured training program containing 4 first-level indicators and 12 second-level indicators was formed. A total of 86 non-endocrinology specialist nurses (working years 2–15 years, average 6.8 ± 3.2 years; professional titles: 32 nurses, 38 senior nurses, 16 charge nurses) were selected by stratified random sampling and underwent a 4-week training. The training included theoretical lectures, scenario simulations, case studies, and bedside practical operations. After the training, the effects were comprehensively evaluated through theoretical assessments, operational scores, and patients’ blood glucose monitoring data. Result: After the training, the standardized rate of perioperative blood glucose monitoring among nurses increased from 54.2% to 91.7% (p < 0.001), the qualified rate of standardized insulin injection increased from 58.3% to 91.8% (p < 0.001), and the average score of theoretical assessment increased from 62.5±8.3 points to 89.6 ± 5.7 points (p < 0.001). The preoperative blood glucose compliance rate of patients increased from 54.3% to 76.8% (p < 0.001), the 24-hour blood glucose fluctuation range after surgery decreased from (5.8 ± 1.6) mmol/L to (2.9 ± 0.9) mmol/L (p < 0.001), and the incidence of perioperative hypoglycemic events decreased from 8.0% to 2.0% (p = 0.008). Multivariate analysis showed that nurses with a training duration of ≥ 4 weeks had a 68% lower risk of postoperative infection in the patients under their charge (OR = 0.32,95% CI 0.13–0.79). Conclusion: Structured training significantly enhances the perioperative blood glucose management ability of non-endocrinology specialist nurses and improves the clinical outcomes of patients. It is suggested that such training be incorporated into the compulsory content of nurses’ continuing education, a regional homogeneous training and certification system be established, and patient-participatory evaluation tools and digital training platforms be developed to promote the continuous improvement of perioperative nursing quality.