Abstract: To tackle the missing of multi-label load identification information,a power load characteristic analysis and intelligent identification method based on SDP information fusion is proposed in this paper. For the problems of modal aliasing and residual auxiliary noise,CEEMD decomposition is used to extract the periodic signal of the current, which improves the robustness of signal decomposition and reduces the reconstruction error. A load fusion characteristic analysis method based on SDP is proposed to address the issue of missing information in feature extraction, which improves the completeness of feature information. On this basis,the load identification method of SDP-YOLOv5 is proposed,and the load intelligent identification model of SDP-YOLOv5 is established. The experimental studies show that the load recognition accuracy of the proposed method reaches 98%,which ensures the non-invasive load monitoring level.