Objective Guided by the Evidence Implementation Model developed by the Fudan University Centre for Evidence-Based Nursing，this study aimed to apply the best evidence for glycemic management in critically ill patients and evaluate its effectiveness. Methods A four-phase approach—comprising evidence retrieval，baseline audit，practice innovation，and outcome evaluation—was employed to implement best practice in glycemic management in the ICU of a tertiary class A hospital in Zhejiang Province. Outcomes were compared before（June-July 2024） and after（October-November 2024） the evidence application，focusing on healthcare providers’ compliance with audit criteria，the proportion of patients within the time in target blood glucose range，glucose variability，and the incidence of hypoglycemia. Results The study included 8 physicians，28 nurses，and 60 patients（30 before and 30 after the intervention）. After the evidence practice，there was a significant improvement in the execution rates of audit criteria among clinicians，with 12 audit indicators showing statistically significant differences compared to baseline（P<0.05）. System-level audit criteria（items 10，24） achieved 100% compliance. The differences in the average proportion of patients within the time in target blood glucose range and glucose variability before and after the evidence-based practice were statistically significant（P=0.026，0.027）. However，the difference in the incidence of hypoglycemia was not statistically significant（P=0.667）. Conclusion The evidence-based practice for blood glucose management effectively standardized the behaviors of healthcare providers，improved the proportion of critically ill patients within the target blood glucose range，reduced glucose variability and the incidence of hypoglycemia.