AIMS: To examine the effects of fluctuations in fasting blood glucose (FBG) levels on left ventricular function in patients with T2DM and coronary microcirculation dysfunction (CMD). METHODS: A total of 290 patients with T2DM who received glucose-lowering therapy during hospitalization and were subsequently followed up for 18 months at the First Affiliated Hospital of Harbin Medical University, were enrolled in this study. 135 were diagnosed with CMD and were assigned to the CMD group, whereas 155 patients without CMD were allocated to the non-CMD group. The fasting blood glucose coefficient of variation (FBG-CV) was calculated for all participants. The CMD group was further stratified into three subgroups based on their FBG-CV values: CMD1 (FBG-CV > 25%), CMD2 (FBG-CV 15% ~ 25%), and CMD3 (FBG-CV < 15%). The left ventricular function, assessed by left ventricular ejection fraction (LVEF) and the E/e' ratio, was compared within each group before and after the follow-up period. This study was registered in the Chinese Clinical Trial Register, ChiCTR-ORC-16009800. RESULTS: After the end of follow-up, the E/e' ratio in CMD was significantly higher than that in CMD and CMD (14.35 vs 8.57; p < 0.01; 14.35 vs 6.61; p < 0.01), and the E/e' ratio in CMD was significantly higher than that in CMD (8.57 vs 6.61; p < 0.01). Compared to the baseline measurements, the E/e' ratio in CMD showed a significant increase after an average 17.8 months of follow up (14.35 vs 8.44; p < 0.001). We found elevated E/e' ratio was associated with an increased FBG-CV level (odds ratio [OR]: 2.571; 95% CI 1.819-3.634; p < 0.001). In multivariate logistic analysis, course of diabetes (OR:1.062; 1.016-1.11; P = 0.007) and CMD (OR:2.231; 1.303-3.819; P = 0.003), were significantly associated with elevated E/e' ratio, while oral stains drugs (OR = 0.412 95% CI 0.237-0.715; P = 0.002) and insulin injections (OR = 0.536 95% CI 0.311-0.924; P = 0.025) behaved as a protective factor. CONCLUSIONS: Our study clarified the association between FBG-CV levels and the E/e' ratio in a prospective cohort study. In T2DM patients with CMD, FBG-CV > 25% may adversely affect left ventricular diastolic function, whereas an optimal FBG-CV is considered to be less than 15%.