There are three kinds of blood vessel systems in liver and the localization of each intrahepatic vessel system is one of the most important issues for liver surgery. These different kinds of vessels are typically visible in each phase of contrast enhanced liver CT Data. In order to merge the different information into one image, image registration and data fusion must be adopted. In this paper, image registration and data fusion for different phases of contrast enhanced liver CT data was studied. Firstly, a method based on joint histogram was proposed to register the intrahepatic vessels. This method was evaluated by applied on two data sets, and it was shown that this method outperformed the method of mutual information in terms of speed. Then, some optimization algorithms, such as Powell, Brent, Golden Section and the second-degree interpolation algorithms, were applied. Finally, the post-registration images were fused with a wavelets- based scheme and the different information was merged into a new image. The experiment results demonstrated the effectiveness of the fusion scheme.