Introduction: This study applied the Auto-EEG-Brain AGE model to evaluate Brain Age (BA) and Brain Age Gap Estimate (Brain AGE) in children with Self-limited epilepsy with centrotemporal spikes (SeLECTS) and to examine their associations with the development of Electrical status epilepticus during sleep (ESES), aiming to identify quantitative biomarkers for risk stratification. Objectives: To characterize regional and stage-specific Brain AGE alterations across disease phases and to develop and validate a predictive model for ESES occurrence. Methods: A total of 453 children with SeLECTS spectrum disorders were included. Brain AGE was estimated from EEG data with bias correction. Group differences were assessed using linear mixedeffects models and propensity score matching. A Cox proportional hazards model was used to construct a risk score and nomogram. Model performance was evaluated by timedependent ROC curves, C-index, calibration, Kaplan–Meier analysis, and decision curve analysis. Results: The BA–chronological age correlation was lower in SeLECTS patients than in controls. Brain AGE showed significant regional gradients, with the greatest positive deviation in the central region and bidirectional abnormalities in the temporal region. Patients who developed ESES exhibited stage-dependent Brain AGE changes with a decreasing trend over disease progression. The Brain AGE–based model demonstrated moderate predictive performance. Conclusions: Brain AGE sensitively captures abnormal neurodevelopmental patterns and dynamic alterations associated with SeLECTS and ESES. The integrated risk prediction model incorporating Brain AGE and clinical variables shows potential clinical utility for early identification and individualized management of high-risk patients.