Most patients with schizophrenia have experienced a period called clinical high risk (CHR) preceding the first episode of psychosis.Early identification of CHR and timely medication or psychological intervention may reduce conversion to psychosis.By exploring biomarkers that predict the onset of psychosis and improving the accuracy of predicting the prognosis of CHR state, it is possible to take a more active intervention measure.As a fast and economical neurophysiological test, event-related potential (ERP) may reflect the perception process, pre-attention process, and attention distribution of the cognitive processes.And it holds promise for becoming objective indices in predicting the clinical outcomes of the CHR patients.This paper reviews the current studies on different ERP components in the CHR population and their performance as predictors of clinical outcomes.The results show that among ERP abnormalities, P300 and MMN amplitude reductions appear to be more reliable than others, which may indicate that distinct components reflect different stages of the disease.However, as a physiological index in the CHR group, aberrant ERP lacks certain specificity.The algorithm analysis combining different ERP components or combining components with symptoms may make the test more specific in the future.