In order to solve the defects such as low matching accuracy induced by lots of wrong matching when facing the big geometric transformation in current image registration algorithm, the precise dynamic image registration algorithm based on weighted phase coherence coupled modified image transformation matching was proposed in this paper. Firstly, the key points of image were detected by SIFT mechanism, then the weighted factor was embedded to define the phase consistency for purification of the key points to eliminate the wrong features as well as instable features points. The angular distance was designed to substitute the neighbor features for improving the image transformation matching technology. Then the mapping relationship between initial matching feature points and accuracy matching feature points was used to correct the initial matching. Finally, the corrected matching relationship was treated by the improved image transformation matching technology. The simulation results showed that this algorithm had stronger robustness and higher matching accuracy. In conclusion, the improved algorithm could enhance the matching accuracy of images at relatively large geometric transformation degree.