The invention discloses a coal dressing system and method based on artificial intelligence image recognition. A conveying belt is arranged, coal materials move along with the conveying belt in a single-row mode, in the moving process, coal ash on the coal materials is washed by a spray gun, part of the solid surface is exposed, a camera device shoots and sends shot images to a control module in real time, gangue and coal briquettes are recognized in real time through an artificial intelligence image recognition technology, then a flow guide sorting mechanism on the downstream of the conveyingbelt is controlled to act, and the gangue and the coal briquettes are separated. The coal ash on the coal materials is washed by arranging the spray gun, so that the visual discrimination degree of the coal and the gangue is increased; then the gangue and the coal briquettes are recognized by applying the artificial intelligence visual image recognition technology of deep learning, so that the accuracy of machine vision recognition can be greatly improved, the sorting precision is high, the equipment cost is low, the processing particle size is wide, the processing capacity is high, most of manual work can be replaced, the labor cost is saved, the coal mine pollution and water consumption are reduced, and the coal loss is reduced; and the system and method are beneficial to being deployedunderground, the gangue well outlet amount is effectively reduced, and environmental protection is facilitated.