The invention discloses an image classification method based on a dense connection residual network, and the method comprises the steps: carrying out the data preprocessing of an input image, and expanding the data volume of the image; inputting the image into a dense connection residual error network to extract image features; and finally, inputting the extracted image features into a Softmax classifier to obtain an image classification result. The method overcomes the defects that a traditional method in an image classification task is poor in effect, a common deep learning method is too large in calculation amount, and model training and operation have too high requirements for hardware equipment, and by combining the advantages of a residual error network and a dense connection network, the high recognition rate is achieved, and meanwhile few calculation resources are occupied.