The invention relates to a wheat hardness prediction method based on near infrared hyperspectral image analysis, and belongs to the technical field of wheat hardness detection. The method mainly comprises the following steps: automatically collecting a near infrared hyperspectral image of kernels, distinguishing effective kernel targets, dividing a sub-image of a single intact kernel, and determining a long axis direction of the kernel; and carrying out digital morphological processing on the whole kernel, so as to obtain a gray sub-image capable of representing the hardness of the single kernel. Aiming at spectroscopic data of the sub-images, the specific hardness value of the tested kernel can be predicted by kernel hardness prediction software; and automatic nondestructive testing of the wheat hardness is realized.