The invention relates to a visual detection method of longitudinal tear of a conveying belt, which belongs to the technical field of on-line monitoring of coal mine automation. The visual detection method is characterized in that in the condition of dim light or no light, a pyroelectric detector is used for detecting the temperature variation of the conveying belt due to the longitudinal tear, an infrared CCD (Charge-coupled Device) camera is controlled to take a photo, an obtained image is processed, and then an accident of the conveying belt can be accurately judged. According to the visual detection method, the simultaneous working mode of timing photo taking and triggering photo taking is adopted, omitted judgment caused by failure of the pyroelectric detector is prevented, and great energy of the infrared CCD camera, which is consumed by continuously taking photos for a long time, is saved. Additionally, negative effects that the service life of the infrared CCD camera is shortened and the working efficiency is reduced are avoided, wherein the negative effects are caused by the fact that the infrared CCD camera is continuously operated for a long time. A sharp image can be obtained by the infrared CCD camera when the conveying belt is moved at high speed, the obtained image is a grayscale image and is easily processed at a later period, and the position and the degree of the longitudinal tear can be visually reflected through finally processed information. The visual detection method has the advantages of high precision and reliable performance.