An oral cavity CBCT image mandibular neural tube automatic recognition method based on Mask RCNN comprises the following steps: collecting oral cavity CBCT image data, preprocessing a coronal diagramin the oral cavity CBCT image data, removing an image which does not display a mandibular neural hole in the coronal diagram, performing format compression on the image, and manually aiming at the mandibular neural hole; generating a sighting frame, positioning the rectangular frame, obtaining a binary mask mask of the target instance, and establishing a neural network model; training a model; andrecognizing and displaying the mandibular nerve hole through the trained model. According to the method, the Mask RCNN is used for automatically identifying the mandibular neural tube in the CBCT image. Machine recognition is used for replacing manual recognition of coronal diagram tooth nerve holes, so that the labor cost is saved, the stability of the generated mandibular nerve tube track is improved, and the method has good recognition speed and precision.