The embodiment of the invention provides a container number identification method and device. The method comprises the steps of collecting first image data of a container door; processing the first image data by using a character detection algorithm based on a convolutional neural network to determine a character area; acquiring graphic features of the container door, and performing feature analysis on the graphic features in the character area to determine a container number area of the container; conducting character recognition on the container number area, and outputting the container number of the container; wherein the graphic features comprise the contour and height of the container door, the character density of the container door and the check bit of the container number; in thisway, using the mobile terminal equipment for collecting container image data, and lowering the requirements for equipment and sites; identifying the image data by utilizing a character detection algorithm based on a convolutional neural network to ensure the success rate of character detection in the natural environment; making characteristic analysis by utilizing graphic characteristics of the container door to increase container number recognition rate.