Prognostication after moderate-to-severe traumatic brain injury (TBI) remains challenging in the intensive care unit (ICU) despite the existence of well-validated online prognostication tools. Changes in brain activity related to TBI can be measured using electroencephalography (EEG), making it a potentially interesting diagnostic tool to refine prognostication. The primary objective of this systematic review was to evaluate the literature concerning the prognostic value of EEG among patients with TBI in the ICU. Five databases were searched from inception until August 13, 2024. The search identified 1492 unique records. Eventually, 27 manuscripts met the inclusion criteria (>18 years old, Glasgow Coma Scale ≤12, EEG performed in the ICU). The QUIPS (QUality In Prognostic Studies) and PROBAST (Prediction model Risk Of Bias ASsessment Tool) tools were used to assess the study quality and bias. Due to high heterogeneity in EEG feature and outcome definitions and a lack of correction for confounding factors, all studies had a moderate-to-high risk of bias. Nonetheless, specific EEG features (identified through visual and quantitative EEG, EEG reactivity, and machine learning techniques) were found to be predictive of neurological outcomes up to 1.5 years after TBI. While epileptiform discharges and seizures were not consistently associated with outcomes, a higher alpha variability, a more continuous EEG, present EEG reactivity, and present EEG sleep features were predictive of better outcomes. The combination of EEG features with clinical parameters demonstrated improved predictive performance compared with models using standard clinical parameters alone. Still, the EEG features described and their potential additional value in outcome prediction after TBI merit further investigation.