The invention discloses a concurrent neural network-based file fragment classification method and device. The method comprises the following steps of: firstly extracting file contents in file fragments; converting the extracted file contents into hexadecimal first character string sequences; taking the character string sequences as input information of a concurrent neural network deep learning algorithm; and finally inputting the first character string sequences into a pre-constructed trained concurrent neural network algorithm-based classification model, and judging types of the file fragments by utilizing the classification model. According to the method, through a learning process of carrying out automatic feature extraction on feature vectors of fragments on the basis of a concurrent neural network classification algorithm, the types of file fragments can be classified.