The invention discloses a device for on-line detection and classification of fiber tripping defects of chemical fiber spindles. The device comprises a spindle conveying device, an image acquisition device and a defect elimination module which are positioned on the spindle conveying device, and a good product waiting area and a bad product waiting area which are connected with the spindle conveyingdevice; the defect elimination module comprises an image processing and decision module, a quality statistical evaluation module and a defect control module; the image acquisition device comprises closed dark boxes, LED light source groups, CCD camera groups, image acquisition cards, digital signal processors and memories; the closed dark boxes are arranged above and below the silk ingot conveying device; and the LED light source groups, the CCD camera groups, the image acquisition cards, the digital signal processors and the memories are arranged in the closed dark boxes. The device has thefollowing beneficial effects that collected silk ingot images are processed by combining machine vision and an image processing technology, the interference number and morphological characteristics ofthe tripping fibers and interference fibers can be simultaneously detected, the fiber ingots with the fiber tripping defects can be detected by using detected fiber tripping information, and a sourcecan be traced by using statistical interference information.