Efficient registration of real-time image and standard image is one of the key steps to ensure the efficiency of cigarette package quality inspection. According to the structure and printing process of cigarette package, a single-angle region registration method is proposed. The method is divided into two steps: rough registration and accurate registration. Rough registration uses the boundary features of cigarette labels to adjust the posture and position of cigarette labels in the image, and removes redundant pixels in the image. The accurate registration uses the homography matrix between the real-time image and the standard image to complete the registration. The experimental results show that the single-angle region registration algorithm is superior than scale invariant feature transformation, ORB, and AKAZE algorithms in registration speed and effect, and can meet the real-time requirements of cigarette package quality detection.