Abstract: In view of problems of low efficiency, environmental pollution and high cost existed in traditional coal and gangue separation method, a separation system of coal and gangue based on X-ray and machine vision was designed. The system uses machine vision to obtain visual images which were identified through image processing algorithm to get the thickness of coal and gangue. Then X-ray was used to scan and obtain coal and gangue images which were identified through the image processing algorithm to get the gray value. Finally, the gray value was combined with thickness of coal and gangue to draw the sorting threshold, so as to separate the gangue from the coal. The test result shows that the system can recognize coal and gangue effectively through the threshold with high recognition rate, and recognition precision meets requirements of industrial production.