The invention relates to an electronic device excess article detection and classification method based on nonnegative tensor decomposition, relating to an excess article detection and classification method of an electronic device and belonging to the technical field of equipment safety detection and protection. The method solves the problem that the detection performance cannot be improved since aural signals and acceleration signals cannot be fully utilized in the existing large and complicated electronic device excess article detection technology. Aiming at the characteristics of the collected aural signals and acceleration signals and the advantages of nonnegative tensor decomposition method for processing data-based multidimensional signals, the nonnegative tensor decomposition method is selected for synchronously processing the aural signals and acceleration signals reflecting characteristics of excess article and the type of excess article so as to directly obtain the characteristic vector of the excess article and achieve the purposes of detection and classification of excess article.