This study aimed to investigate the relationships between visit-to-visit variability in HbA1c and fasting plasma glucose (FPG) and atherosclerotic markers in persons with type 2 diabetes mellitus (T2DM), employing epidemiological analyses and Mendelian randomization (MR) techniques. Between January 2001 and December 2020, 2290 persons with T2DM (1307 men and 983 women) who underwent evaluation for atherosclerosis screening were enrolled in a management program at a medical center in Taiwan. To assess potential causal relationships among the coefficient of variation (CV) of HbA1c (HbA1c-CV), FPG-CV, and markers of atherosclerosis, we conducted a 2-stage multivariable regression analysis using 22 and 21 single nucleotide polymorphisms as instrumental variables for HbA1c and FPG, respectively. A total of 2290 individuals with T2DM were included, with a mean age of 61.79 years, and 57% were male. In the epidemiological analysis, after multivariable adjustment, the odds ratios for an ankle-brachial index (ABI) < 0.9 in the third and fourth quartiles compared with the first quartile were 3.15 (95% confidence interval [CI]: 1.80-5.50) and 4.29 (95% CI: 2.44-7.56), respectively, for FPG-CV, and 1.67 (95% CI: 1.06-2.62) and 1.65 (95% CI: 1.01-2.70), respectively, for HbA1c-CV. In the MR analysis, comparing the fourth quartile with the first quartile, the odds ratios for ABI < 0.9 and brachial-ankle pulse wave velocity (baPWV) ≥ 1800 cm/s were 2.78 (95% CI: 1.84-4.19) and 1.58 (95% CI: 1.23-2.02), respectively, for the unweighted genetic risk score predicting FPG-CV. Epidemiological and MR analyses provide evidence that FPG variability is significantly associated with atherosclerosis, as indicated by the ABI.