In view of the fault detection of computer equipment, a detection method based on embedded technology is proposed. Acquire operation data of computer equipment by utilize a wireless sensor, constructing a MobileNet-SSD network model, realizing that improvement of the constructed network model through the fusion of a convolution lay and a Batch Normalization layer, ensuring that model information before pruning is reserved by adopting a pruning fine-tuning optimization method, transplanting the pruned model to embedded equipment, The three fault characteristic quantities are used as the input of the model to complete the operation fault detection of the computer equipment. The experimental results show that the method can complete the extraction of computer equipment fault features and obtain equipment fault detection results; the convergence rate is fast, and the training loss is less than 0.01; the compression of fault detection model can be realized, the number of parameters decreases significantly, and the fluctuation of mAP index is small.