Objective To investigate the correlation between glycemic variability metrics and the risk of diabetic foot （DF） in patients with type 2 diabetes mellitus （T2DM） utilizing flash glucose monitoring （FGM） technology. Methods A retrospective analysis was conducted on 233 hospitalized patients with T2DM， with or without DF， who were treated in the Department of Endocrinology at Lianyungang First People's Hospital from January 2021 to May 2022 and monitored using FGM. Patients were categorized into a non?DF group （n = 147） and a DF group （n = 86） based on the presence of DF. The study compared general clinical characteristics， biochemical parameters， and glycemic variability metrics between the two groups and performed subgroup analyses. Binary logistic regression was employed to identify factors associated with the risk of DF， while receiver operating characteristic （ROC） curves were utilized to assess the predictive value of glycemic variability metrics for DF. Results Compared with the non?DF group， patients in the DF group exhibited significantly longer disease duration， higher body mass index （BMI）， glycated hemoglobin （HbA_1c）， urinary albumin?to?creatinine ratio （UACR）， alanine aminotransferase （ALT）， serum uric acid （SUA）， mean amplitude of glycemic excursions （MAGE）， coefficient of variation （CV）， mean of daily differences （MODD）， and mean blood glucose （MBG）， but lower fasting C?peptide （FCP）， fasting insulin （FINS）， high?density lipoprotein cholesterol （HDL?C）， and time in range （TIR）， with statistically significant differences （P < 0.05）. Subgroup analysis revealed that TIR was associated with the incidence of DF and diabetic retinopathy （DR）. Binary logistic regression analysis identified HbA_1c， MAGE， MODD， and MBG as risk factors for DF， while TIR was a protective factor （P < 0.05）. ROC curve analysis demonstrated that the area under the curve （AUC） for predicting DF using HbA_1c， TIR， MAGE， MODD， MBG， and their combination were 0.646， 0.850， 0.868， 0.764， 0.619， and 0.967， respectively， indicating superior performance of the combined prediction model. Conclusions HbA_1c， TIR， MAGE， MODD， and MBG are critical factors associated with the development of DF in patients with T2DM. Targeted early interventions aimed at optimizing these glycemic variability indicators may effectively reduce the incidence of DF.