ObjectiveTo propose and evaluate the clinical utility of a new nutrition screening algorithm, NutriPal, to detect the degree of nutritional risk in patients with incurable cancer receiving palliative care.MethodsIt is a prospective cohort conducted in an oncology palliative care unit. The NutriPal algorithm was used in a three‐step process: (i) administration of the Patient‐Generated Subjective Global Assessment short form; (ii) calculation of the Glasgow Prognostic Score; and (iii) application of the algorithm to classify patients into four degrees of nutritional risk. The higher the degrees of NutriPal, the worse the nutritional risk, comparing nutritional measures, laboratory data, and overall survival (OS).ResultsThe study included 451 patients that were classified using the NutriPal. They were allocated to the degrees: 1 (31.26%), 2 (27.49%), 3 (21.73%), and 4 (19.71%). Statistically significant differences were found in most of the nutritional and laboratory parameters and in OS with each increment in the NutriPal degrees, and OS was reduced (log‐rank <0.001). In addition, NutriPal was able to predict a 120‐day mortality: there was a significantly higher risk of death in the patients classified as degrees 4 (hazard ratio [HR], 3.03; 95% confidence interval [95% CI], 2.18–4.19), 3 (HR, 2.01; 95% CI, 1.46–2.78), and 2 (HR, 1.42; 95% CI; 1.04–1.95) than in those classified as degree 1. It also showed good predictive accuracy (concordance statistic, 0.76).ConclusionThe NutriPal is associated to nutritional and laboratory parameters and can predict survival. It could therefore be incorporated into clinical practice for patients with incurable cancer receiving palliative care.