In order to solve the problems of long computation time and low fault detection accuracy in current microgrid protection schemes, a microgrid fault detection model based on hybrid deep learning is proposed. The feature advancer is used to mine the signal information of power data, and the deep convolutional neural network is used to effectively extract the feature information of power fault data, and the AdaBoost classifier is used to classify the fault. Experimental results show that, contrary to the convolutional neural network (CNN) and AlexNet, the proposed hybrid deep learning detection model has higher training performance; Compared with SVM, LR, CNN and AlexNet models, the proposed hybrid deep learning model has better comprehensive index performance, and the fault detection accuracy can reach 98%.