The invention discloses a multimode characteristic information fusing and predicting method of depression suicide behavior. The multimode characteristic information fusing and predicting method comprises the steps of collecting EEG signals, during data collection, using one-sided mastoid as a
reference electrode, using opposite side mastoid as a
recording electrode, besides, performing three-dimensional psychology pain measurement on testers, and performing neuropsychology behavioral indicator judgment on an simulated
balloon risk task; preprocessing data, detecting EEG
signal synchronicity onthe preprocessed
EEG data through calculating PLV value between brain electrodes, and performing
brain network construction through the PLV value; and by a
pattern recognition method, performing classification on depression patient samples having high-risk suicide behavior and depression patient samples having low-risk suicide behavior. According to the multimode characteristic information fusingand predicting method disclosed by the invention, classification accuracy can be effectively increased, prediction of high suicide risk of
major depressive disorder is objective, and the occurrence of a phenomenon of the suicide behavior is effectively reduced.