The invention discloses a bridge modal parameter intelligent updating method based on a cross-modal confidence criterion (CMAC) matrix, and the method comprises the steps: taking a CMAC matrix as a basis, combining an adaptive convolution operation layer and a full-connection classification layer to form a modal spectrum response intelligent extraction neural network, and carrying out the classification reconstruction of the CMAC matrix; extracting a bridge structure physical modal spectrum response interval; furthermore, establishing an agent model of bridge vibration data power spectral density and modal information intensity based on the initial modal spectral response interval and a finite element model modal parameter theoretical value, so that a maximum modal information spectral response interval is determined, and bridge structure modal parameter identification is carried out accordingly. According to the method, the CMAC matrix and the adaptive convolutional neural network are combined to carry out intelligent analysis and identification of the structural modal parameters, the network training efficiency is high, the response of the weak excitation modal can be well extracted, and the method can be applied to bridge structure health state monitoring to carry out automatic updating of the modal parameters.