Background To investigate the impact of leukocyte glycemic index (LGI) on the prognosis of sepsis patients with new-onset atrial fibrillation (NOAF).  Methods This retrospective cohort study utilized data from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database to evaluate adult patients with sepsis and new-onset atrial fibrillation (NOAF). The leukocyte glycemic index (LGI) was derived from the initial white blood cell count and blood glucose values recorded within 24 hours of intensive care unit admission, with log-transformation applied for analysis. Participants were divided into three groups according to LGI tertiles, employing 30-day all-cause mortality as the primary endpoint. Kaplan-Meier survival curves were constructed, with log-rank tests used to assess intergroup differences. Cox proportional hazards regression models and restricted cubic spline (RCS) analyses were applied to investigate the association and nonlinear relationship between LGI and the endpoint, complemented by subgroup analyses and interaction tests.  Results A total of 691 patients were enrolled, exhibiting a 30-day all-cause mortality rate of 45.59%. The cohort had a median age of 72.93 years and comprised 418 males. Kaplan-Meier survival analysis revealed significantly higher mortality rates among patients with elevated LGI levels (p < 0.05). Multivariable Cox proportional hazards models demonstrated that individuals in the highest LGI tertile faced a substantially increased risk of 30-day mortality relative to the reference tertile. Restricted cubic spline analysis (RCS) identified a significant U-shaped nonlinear association between log-transformed LGI and mortality, with an inflection point at 4.191; beyond this threshold, each one-unit increase in log-transformed LGI conferred a 36% higher mortality risk (HR = 1.36, 95%CI: 1.11–1.67, p = 0.003).  Conclusions In patients with sepsis and new-onset atrial fibrillation, the leukocyte glycemic index exhibits a significant U-shaped nonlinear relationship with 30-day all-cause mortality. As a readily obtainable biomarker that integrates inflammatory intensity and glycemic dysregulation, LGI holds promise for enhancing early risk stratification and prognostic evaluation in this high-risk cohort. Prospective investigations are needed to corroborate these results and elucidate the underlying pathophysiological mechanisms.