Abstract: In digital image processing, edge detection is an important content.Commonly used edge detection methods, such as the method of detection the maximum gradient, the method of detection the second derivative of the zero-crossing point, statistical-based method and wavelet multiscale edge detection method, etc., there is difficulty to determine the parameters of reasonable threshold.The Edgeflow method synthesizes the brightness, texture and phase characteristics of a variety of image information to meet the opposite direction of the edge of streams to determine the location of the edge of an object, it solved the problem that the threshold value of the traditional image segmentation algorithm based on the edge is difficult to determine.This paper discusses the principle of image segmentation algorithm based on edge flow, adjusts the algorithm and designs a predictive coding model to achieve.The results show that: Edgeflow method adjusts parameter less and detection effect is better.