Background <p>This study aimed to explore the relationship between glucose variability (GV) and echocardiographic parameters and to investigate the mediating effect of echocardiographic parameters in the association of GV with all-cause and cardiovascular mortality.</p> Methods <p>In this retrospective cohort study, we acquired data from the electronic health records of individuals with type 2 diabetes mellitus (T2DM) who visited a medical center during 2001–2020 (Jan to Oct) with follow-up until December 31, 2021. Cox proportional hazards models were conducted to explore the relationship of GV and echocardiographic parameters and mortality. We employed mediation analysis to assess the role of echocardiographic parameters in linking GV to mortality.</p> Results <p>There were 3,996 patients with T2DM. Both variability in HbA1c and fasting plasma glucose (FPG) were significantly linked with mortality. FPG variability was correlated with left ventricular mass index (LVMI), left ventricular mass (LVM), e’, and tricuspid regurgitation (TR) velocity according to multivariable regression analyses. At a mean follow-up of 5.2 years, FPG variability, LVM, LVMI, s’, and TR velocity were independently linked with all-cause and expanded cardiovascular disease (CVD) mortality. LVM, LVMI, s’, and TR velocity mediated 3.04%, 3.75%, 2.19%, and 2.71% of the association between FPG variability and mortality from any cause, respectively; and s’ and TR velocity mediated 4.33% and 2.72% of the effect of FPG variability on expanded CVD mortality, respectively. All the tested echocardiographic parameters were not significant mediators of the relationship of HbA1c-CV with mortality from any causes and expanded CVD.</p> Conclusions <p>Our study found a weak mediating effect of the echocardiographic parameters of cardiac structure and function in the association between FPG variability and mortality, but not in the association between HbA1c variability and mortality.</p>