It has been widely used in computer vision, military investigation, medical care and some other fields. The widely used seamline removal methods include overlappingimage method, wavelet transform method and hardcorrection method. An improved method based on the correctionratio of average gray value of the images on the two sides of the seamline was proposed. This method calculated the rate of the corrected value and the former value. Then the ratio gradually changed from the seamline to the boundary so that the seamline was removed. The experiments have shown that the method proposed above is effective to eliminate seamlines, especially when the difference of the gray level is big.