This paper presents a novel background model called Grayscale Arrange Pairs(GAP) that excels at analyzing scenes under varying illumination condition. The main concept of the proposed method is to use multiple point pairs that each point pair has a specific intensity relationship as a background model. If the scene contains no moving object, changes in the intensity difference between each pixel of the pair will be less than each pixel’s individual intensity even in varying illumination conditions. In contrast to the existing approaches, our proposed method focuses more attention onto the correlations between pixels than on the history of any given pixel. Furthermore, We show how the image processing time for modeling can be reduced and present experimental results comparing GAP to existing event detection methods, demonstrating that superior event detection with high precision and recall rates is achieved by GAP.