Background <p>Diabetic retinopathy (DR) is one of the most important microvascular complications of diabetes mellitus and remains a major cause of preventable blindness worldwide. Insulin resistance plays a pivotal role in the development of diabetes and its associated complications. The triglyceride–glucose (TyG) index, a surrogate marker of insulin resistance derived from fasting plasma glucose and triglyceride levels, has gained attention as a simple and practical metabolic marker. However, data linking the TyG index with diabetic retinopathy, particularly in South Asian populations, are still limited.</p> Methods <p>A hospital-based cross-sectional study was conducted at Bir Hospital, Kathmandu, Nepal, between April 2023 and October 2024. Adult patients with diabetes mellitus, excluding those with gestational diabetes, acute illness, non-diabetic causes of hypertriglyceridaemia, conditions affecting lipid or glucose metabolism, prior retinal interventions, or non-diabetic retinal diseases, underwent comprehensive clinical, biochemical, and ophthalmic evaluations. The triglyceride–glucose (TyG) index was calculated as ln [fasting triglycerides (mg/dL) × fasting plasma glucose (mg/dL) / 2], and participants were categorized into quartiles. Diabetic retinopathy was graded using internationally accepted clinical severity scales. Associations were assessed using univariate and multivariate logistic regression, and ROC curve analysis evaluated the discriminatory performance of the TyG index.</p> Results <p>Among 83 patients with diabetes mellitus (median age 56 years [IQR 48.5–64], 50.6% male, median diabetes duration 6 years [IQR 3–12]), the median triglyceride-glucose (TyG) index was 9.98 (IQR 9.485–10.445). Diabetic retinopathy (DR) prevalence was 48.2%, increasing across TyG quartiles from 9.5% in Q1 to 85.7% in Q4 (<i>p</i> &lt; 0.001), with a significant trend in DR severity (<i>p</i> &lt; 0.001). TyG correlated moderately with HbA1c (Spearman <i>r</i> = 0.54, <i>p</i> &lt; 0.001). In multivariable logistic regression adjusted for age, total cholesterol, and HDL, the odds ratio for DR per standard deviation increase in TyG was 4.00 (95% CI 1.75–9.16, <i>p</i> = 0.001); for Q4 vs. Q1, 24.15 (95% CI 3.15–185, <i>p</i> = 0.002). Findings were robust in sensitivity analyses excluding insulin users (OR 4.35, 95% CI 1.40–13.5, <i>p</i> = 0.011) and outliers (OR 3.80, 95% CI 1.60-9.00, <i>p</i> = 0.002), and stratified by diabetes duration (≥ 5 years: OR 5.20, 95% CI 1.80–15.0, <i>p</i> = 0.002) and sex (males: OR 3.80, 95% CI 1.20–12.0, <i>p</i> = 0.023; females: OR 4.50, 95% CI 1.30–15.5, <i>p</i> = 0.017). TyG predicted DR with an area under the ROC curve of 0.83 (95% CI 0.75–0.91), comparable to HbA1c (0.77, DeLong <i>p</i> = 0.29); optimal cutoff 9.78 (sensitivity 88%, specificity 67%).</p> Conclusions <p>A higher triglyceride–glucose index was strongly associated with both the presence and severity of diabetic retinopathy. Given its simplicity and reliance on routinely available laboratory parameters, the TyG index may be a useful marker for identifying diabetic patients at increased risk of retinopathy, particularly in resource-limited settings.</p> Clinical trial registration <p>Not applicable. This was a cross-sectional observational study and did not involve a clinical trial requiring registration.</p>