The existing damage identification methods are faced with the problem that the processing capacity of a large number of bridge health monitoring data is limited and the extracted information cannot fully reflect the health status of the bridge. Based on the advantage of deep learning in big data，a method of bridge structure damage location identification based on stacked denoising auto-encoder is proposed. The method is verified by a finite element model of a simply supported beam bridge. The vertical acceleration time history response value at L / 5，2 L / 5，3 L / 5，4 L / 5 of the bridge are extracted. The acceleration matrix (4 x 1) at every moment is established，then the preprocessed matrix is sent to the stacked denoising auto-encoder for feature extraction to complete the model training. Finally，the test samples are sent to the model for classification to complete the task of damage location identification. The results show that the proposed method has the advantages of high positioning accuracy and good anti-noise performance compared with the traditional machine learning method，and has certain reference value in the field of bridge structure damage identification in the future.