In this paper, a kind of cigarette tipping paper detection system based on machine vision is proposed, which uses STM32 as the core controller to control the timing synchronization of the coiler. The cigarette tipping paper on the transition drum is collected by the industrial camera, and the length of the cigarette paper is calculated and measured, and the unqualified cigarette is eliminated. At the same time, the unqualified cigarette tipping paper and detection data are recorded by the guide rail scale. Through practical application, the detection accuracy of the system is better than ±0.05 mm, and the maximum detection speed is 12 000 pieces/min. It can remove unqualified cigarette and reduce the rejection rate of the wrong length of cigarette paper to 0.01%. This paper analyzes the influence factors of error fluctuation through spectrum analysis method, and realizes the evaluation of equipment operation state and cigarette tipping paper quality.