BACKGROUND: This study aimed to identify trajectories of multimorbidity in older adults prior to receiving long-term care benefits and to demonstrate their value in predicting mortality. METHODS: This study included 1,004,924 Korean beneficiaries who completed the National Long-Term Care Insurance (NLTCI) eligibility assessment between 2010 and 2016. Multimorbidity was defined as the coexistence of 2 or more out of 23 chronic diseases related to disability in the 10 years before transitioning to long-term care. Mortality was defined as all-cause deaths after the date of the NLTCI needs assessment. Latent class growth modeling was performed to identify groups that exhibited similar trajectory patterns over time. Sex, age, and long-term care grade were used as covariates. Cox proportional hazards models were used to analyze the mortality rates by trajectories. RESULTS: Three patterns emerged in the multimorbidity trajectory in the 10 years prior to entering the long-term care system: consistently low morbidity ("consistently low"), an abrupt increase in morbidity in less than one year ("catastrophic"), and an increment in morbidity over a longer period ("progressive"). In multiple Cox regression adjusting for covariates, the hazard ratios (95% confidence interval) of 1-year mortality for the catastrophic and progressive groups were 1.38 (1.36-1.39) and 1.43 (1.41-1.45), respectively, compared to the consistently low group. CONCLUSIONS: This study identified distinct trajectories of multimorbidity in older people accessing the long-term care system and demonstrated their prognostic value for the survival of those with long-term care needs. Treatment and management strategies targeting individuals with a high-risk trajectory are warranted.