Abstract: Nowadays some vegetable farmers pick unripe tomatoes and treat them with ethylene to quicken ripeness in China. In order to keep artificial ripening tomato which harming consumer’s health from entering into melon and fruit market, the hardware structure of artificial ripening tomato recognition system was given. The colour parameters RGB (red, green, blue) of transmitted light of tomatoes were obtained through computer vision device, and the RGB values were converted into HIS (hue, intensity, saturation) values. The multilayer feedforward neural networks with genetic algorithm training realized the automated recognition of artificial ripening tomato. The results of test showed that accurate recognition rate of the system was 91.7%, and the method can provide references for further research on recognition of artificial ripening tomato and nature mature tomato.