Background <p>Limited research has explored survival outcome predictions in those with advanced dementia within the Chinese population. A longitudinal cohort study was performed to better characterize the end-of-life trajectories of Chinese individuals with advanced dementia, providing a means of identifying prognostic predictors of 32-month mortality among nursing home residents in Chengdu, China.</p> Methods <p>This study enrolled 186 residents who met the inclusion criteria, with all having advanced dementia. Participants were evaluated at baseline and followed until death or the study’s conclusion. Demographic information, health and medical status, and mortality-related data were assessed. Kaplan-Meier survival analyses and Cox proportional hazards models were applied to approximate survival rates and identify predictive factors.</p> Results <p>The median age of enrolled participants was 83.5 years, with females comprising 60.2% of the cohort. The overall survival rate was 52.0%. After adjusting for all variables, three significant prognostic factors were identified: oral nutritional supplement use (HR: 3.006, 95% CI: 1.593–5.671, <i>p</i> = 0.001), albumin concentration (HR: 0.914, 95% CI: 0.861–0.970, <i>p</i> = 0.003), and memantine use (HR: 0.416, 95% CI: 0.190–0.907, <i>p</i> = 0.028). Survival at 6 months, 1 year, and 2 years was significantly higher in the memantine group versus non-users.</p> Conclusions <p>The present study identified key prognostic factors for mortality in Chinese nursing home residents with advanced dementia. Notably, memantine use emerged as a protective factor, marking the first demonstration of its potential survival benefit in this population. These findings may aid healthcare professionals in optimizing resource allocation and care strategies for individuals at the end of life.</p> Trial registration <p>ChiCTR2100041764 (Chinese Clinical Trial Registry). Registered 4 January 2021.</p>