Addressing the issue of the high concealment of fault features in medical imaging equipment, where a single SVM method struggles to effectively extract deep-level fault features, leading to a low Kappa coefficient in fault classification, a fault classification method for CT medical imaging equipment based on Particle Swarm Optimization (PSO) SVM is proposed. Multiple types of sensors were utilized to monitor the operational status signals of CT medical imaging equipment in real-time. The collected signals were processed through wavelet transform to remove noise and extract valid features. The processed signals were then input into a Deep Belief Network (DBN). The DBN, by stacking Restricted Boltzmann Machines (RBMs) layer by layer and undergoing two stages of unsupervised training and supervised parameter tuning, precisely captured and learned the deep-level fault features of CT medical imaging equipment. The extracted fault features were fed into a multi-classifier based on PSO-SVM, and CT medical imaging equipment fault classification was achieved through the trained model. Experimental results demonstrate that, with the DBN configured with 4 layers and 80, 150, and 80 neurons in the input, hidden, and output layers respectively, the proposed method achieved an F1 score of 0.925, a Kappa coefficient of 0.0895, and a Hamming distance below 0.053. These validation data indicate that the proposed method can accurately classify faults in CT medical imaging equipment, providing powerful technical support for fault diagnosis and maintenance of medical imaging equipment.