The work aims to design the online detection system for tipping paper based on PLC and machine vision, in combination with the actual production requirements, regarding such problems as quality defect of tipping paper occurring after the cigarette production. Line-array cameras were used to collect images. Industrial computer processed the images with the image processing software Vision-Pro and simplified the image processing course with the improved SIFT algorithm. The quality of tipping papers was detected online. Meanwhile, parameter setting, condition monitoring, data recording and other functions were completed. The actual running results indicated that, the percent of pass of the finished cigarettes was improved from 92% to 100%. The reliably operated system can collaborate with the existing system to achieve the online monitoring, which reduces the waste of raw materials and improves the percent of pass of the finished cigarettes.