ObjectiveEvaluating the change trajectories of triglyceride‐glucose (TyG) index calculated after multiple tests in elderly heart failure (HF) patients may have clinical implications for predicting long‐term adverse cardiovascular events (MACEs).MethodsThis retrospective study included 1184 elderly HF (LVEF ≥50%) patients with diabetes admitted to our center between January 2015 and January 2020. Based on the multiple TyG levels detected during the exposure period with annual measurements, three distinct TyG trajectories were determined using latent mixture modeling: low‐stable group (TyG index <8.26, n = 367), medium‐stable group (TyG index 8.26–9.06, n = 613), and high‐increasing group (TyG index >9.06, n = 204). The primary endpoint was the composite outcome of MACEs.ResultsThere were significant increases in the prevalence of several cardiovascular risk factors and conditions, such as male gender, BMI, current smoker, hyperlipidemia, atrial fibrillation, old myocardial infarction, fasting glucose, triglycerides, and uric acid levels, from the Low‐Stable Group to the High‐Increasing Group (all P < 0.05). During a median follow‐up of 29 months (range, 18–46 months), 181 MACEs occurred. Kaplan–Meier analyses curve showed a significantly increased risk of MACEs in the medium‐stable and high‐increasing groups compared to the low‐stable group (HR = 2.528, 95%CI: 1.665–3.838; HR = 2.706, 95%CI: 1.722–4.255, respectively). Furthermore, the rates of heart failure‐related hospitalizations, nonfatal myocardial infarctions and non‐fatal stroke were significantly increased in the medium‐stable and high‐increasing groups. Multivariable Cox regression analyses revealed that age (HR = 1.728), current smoker (HR = 1.385), old myocardial infarction (HR = 1.593), chronic renal disease (HR = 1.682), HbA1c (HR = 1.816), NT‐proBNP (HR = 2.471), TyG trajectory (HR = 2.112) and SGLT2 inhibitors (HR = 0.841) were independently associated with the occurrence of MACEs during follow‐up (P < 0.05).ConclusionsThe TyG trajectory is strongly associated with the risk of MACEs during follow‐up in elderly patients with T2DM, suggesting that TyG trajectories may further optimize risk stratification model of cardiovascular events.