This paper proposes a GCB fault diagnosis model based on adaptive convolutional weights learning module and multi-source data fusion technology. The sound pattern data generated during the operation of the GCB equipment, the base wave voltage spectrograms on both sides of the GCB, and the UHF localized discharge detection spectra are selected as the input data for the fault diagnosis of the GCB equipment; the wavelet transform is performed on the sound pattern data to generate the time-frequency feature maps of the sound pattern; a convolutional neural network is utilized to perform feature extraction on various types of images; the extracted features are used as the input information, and are fed into the feature fusion module for adaptive convolutional weight learning to perform feature fusion; the fused features are fed into a deep neural network to classify the fault diagnosis. The fusion module is used for feature fusion; the fused features are fed into the deep neural network for classification of fault diagnosis. The experimental results show that the method proposed in this paper has a high fault diagnosis accuracy, completeness and precision rate, and has a strong adaptability to complex fault environments.