Objective: Individuals with schizotypal traits can be considered at high-risk for schizophrenia. Studies have shown that individuals with schizotypal traits exhibited neurophysiological abnormalities. However, whether and to what extent could electroencephalogram (EEG) data discriminate individuals with high and low schizotypal traits remained unknown. The present study aimed to examine this issue using a deep learning approach. Method: The resting-state EEG data were collected in 48 individuals with high schizotypal traits and 50 individuals with low schizotypal traits during both eyes-open and eyes-closed conditions. Three EEG datasets were constructed: the eyes-open dataset, the eyes-closed dataset, and the combined dataset. Subsequently, the EEG data of the two groups were classified using the Long and Short-Term Memory Network combined with a one-dimensional Convolutional Neural Network (LSTM-1DCNN) model. Results: The LSTM-1DCNN model demonstrated high accuracy in identifying individuals with schizotypal traits across the eyes-open, eyes-closed and combined datasets, with an accuracy of 94.86%, 94.26%, and 95.30%, respectively. The state of participants’ eyes (open or closed) did not affect the identification accuracy. Conclusion: Individuals with high schizotypal traits exhibited distinct EEG patterns compared to those with low schizotypal traits. EEG data and deep learning algorithm can be employed to identify individuals at risk for schizophrenia.