Aiming at the defects that the threshold should be set artificially, and the detected edge is thick and discontinuous in the traditional edge-detection algorithms based on grey correlation analysis, an adaptive edge-detection algorithm was proposed. After simulating and testing the algorithm, the experimental results indicate that the proposed edge detector can not only get over the above drawbacks of the traditional algorithm effectively, but also adaptively calculate the threshold for different grey image. The extracted edge is more complete, continuous, and fine.