Traumatic brain injury (TBI) is the main cause of death and disability in adults worldwide. Early detection of TBI would be useful for evaluating and designing treatment strategies. Both single predictors from early clinical examination and multiple hospitalization variables/parameters can be used to determine the long-term prognosis of TBI. Predictive models like the IMPACT or CRASH prognosis calculator (based on large sample sizes) can predict mortality and unfavorable outcomes. Moreover, imaging techniques like MRI (Magnetic Resonance Imaging) can also predict consciousness recovery and mental recovery in severe TBI, while biomarkers associated with stress correlate with, and hence can be used to predict, severity and mortality. All predictors have limitations in clinical application. Further studies comparing different predictors and models are required to resolve limitations of current predictors.