Objective: To investigate whether triglyceride-glucose (TyG) index, a marker of insulin resistance, can predict the clinical outcome of patients with locally advanced and advanced non-small cell lung cancer (NSCLC) treated with tislelizumab combined with chemotherapy. Methods: We retrospectively analyzed data from 243 patients who were treated with tislelizumab combined with chemotherapy with locally advanced and advanced NSCLC, dividing them into high and low TyG index groups based on an optimal cut-off value of 9.3, determined using the X-tile software. Cox proportional hazards regression analysis identified independent prognostic factors for overall survival (OS), and a prediction nomogram model was developed based on them. Results: The study cohort included 160 patients with a high TyG index and 83 with a low TyG index. The median progression-free survival and OS in the low TyG index group were significantly longer than in the high TyG index group. The objective response rate was 55.42% in the low TyG index group and 30.63% in the high TyG index group, while the respective disease control rates were 95.18% and 56.88%. Multifactorial regression analysis identified pre-treatment maximum tumour diameter, proportion score for programmed death ligand 1 (PD-L1) positivity, body mass index (BMI), and TyG index as independent prognostic predictors of OS. The nomogram model emphasized the importance of the TyG index (hazard ratio: 2.99, 95% CI: 1.73–5.16, P < 0.001) as a major prognostic predictor, followed by the PD-L1 tumor proportion score, pre-treatment tumour diameter, and BMI. Conclusion: TyG index is a long-term predictor of the efficacy of combined immunotherapy and chemotherapy in patients with locally advanced and advanced NSCLC, and low TyG index has a better prognosis.