Abstract The study aimed to evaluate the relationship between immunonutritional indicators such as the systemic immune-inflammation index (SII), the Naples prognostic score (NPS), nutritional risk index (NRI), serum albumin (ALB), total cholesterol (TC) and all-cause mortality in adult stroke patients. Data were obtained from the National Health and Nutrition Examination Survey (NHANES) databases for 2005–2018. To determine mortality outcomes, participants were matched with National Death Index records until December 31, 2019. Spearman’s correlation analysis and the random survival forest (RSF) were employed to assess the relationships among NPS, NRI, SII, ALB, and TC, and to determine the most predictive indicator for all-cause mortality in stroke patients. For the selected prognostic indicator, Kaplan–Meier survival analysis and Cox proportional hazards regression models were subsequently utilized to evaluate their associations with all-cause mortality in stroke patients. The study included 1076 stroke patients, with a median (IQR) age of 67 (56, 77) years. During a median follow-up of 67 months, a total of 372 (weighted 31%) stroke participants died from all causes. Among the immunonutritional indicators evaluated, NPS had the strongest predictive power for all-cause mortality in stroke patients. The Kaplan–Meier curve and Log-rank test showed that all-cause mortality was higher in the higher NPS group (3–4) compared to the lower NPS group (0–2) (P