An unstructured path recognition and robot vision guidance method based on the fuzzy-rough set is proposed. With the image definition automatic control algorithm based on the self-adaptive surface array charge-coupled device, the images with the best amount of information are obtained. The unstructured path recognition model based on a fuzzy-rough set is built, in which the image target, background and uncertainty area are predefined under the help of the rough set theory and the relative fuzzy connectedness competition mechanism is fused to reclassify the pixels in the uncertain region according to the ambiguity and the robot navigation path is delineated. With this model, the automatic recognition of the unknown and unstructured path areas can be realized, and the gray priori features to recognize the specified path areas can be introduced. The results show that the proposed method has a practical significance for improving the autonomous exploration ability of mobile robots in unstructured environments.