Background Critically ill patients who are hospitalized in a neurological intensive care unit (NICU) are largely susceptible to nutrition risk. However, only a limited number of studies have investigated the applicable assessment tool in NICUs. Methods This was a prospective observational study conducted at a single-center NICU. A total of 140 adult patients who were hospitalized for >24 hours were enrolled. The Nutritional Risk Screening 2002, Nutrition Risk in the Critically ill (NUTRIC), and modified NUTRIC (mNUTRIC) scores were applied for the assessment of nutrition risk. Analyses of multivariable logistic regression were performed by considering a 28-day mortality as the outcome of interest. Results Nutrition risk was commonly identified in NICU patients. Multivariate analysis revealed that age >= 60 years, hospital infection, mechanical ventilation, and high nutrition risk (mNUTRIC score >= 5) independently increased 28-day mortality in NICU patients. For subgroups of patients with a prolonged length of stay, high nutrition risk (mNUTRIC score >= 5) has always been an independent risk factor of 28-day mortality. Both NUTRIC and mNUTRIC scores were able to predict 28-day mortality, with area under the receiver operating characteristic curves of 0.857 (95% CI, 0.786-0.928) and 0.856 (95% CI, 0.786-0.927), respectively. Conclusion The mNUTRIC scoring system is not only a useful tool for nutrition risk assessment but also, and more importantly, it is independently related to the risk of 28-day mortality in NICU patients. Therefore, mNUTRIC scoring is an appropriate tool for nutrition risk assessment and prognosis prediction of NICU patients.