Abstract: As the impact of early albumin infusion on the prognosis of elderly individuals diagnosed with sepsis remains uncertain, this study aimed to investigate this effect in elderly patients with sepsis in the intensive care unit (ICU). We identified the information of elderly patients with sepsis requiring ICU admission from the Medical Information Mart for Intensive Care (MIMIC-IV) database. They were divided into hypoalbuminaemia group and control group, and the primary outcome was 90-day mortality. A multivariate logistic regression model and a multivariate Cox proportional-hazards model were used to analyze the correlation between hypoalbuminaemia and the prognosis. Kaplan-Meier survival curve and log-rank test were performed to analyze the survival outcome. Propensity score matching (PSM) was implemented to determine the precise effect of early albumin infusion on the prognosis of elderly ICU patients with sepsis, and subgroups of patients were identified to explore the factors influencing the relationship. Early hypoalbuminemia was strongly associated with an increased risk of adverse clinical outcomes in elderly patients with sepsis in the ICU. In-hospital mortality (28.6% vs. 19.1%, P < 0.001) and 90-day mortality (48.8% vs. 33.4%, P < 0.001) were both significantly higher in the early hypoalbuminemia group than the control group. PSM analysis showed that early albumin infusion was associated with lower in-hospital mortality and 90-day mortality in elderly patients with sepsis combined with hypoalbuminemia in the ICU. Early infusion of albumin could improve prognosis and reduce in-hospital mortality and 90-day mortality rates.