In recent years biometrics security technology, which certify an individual with human characteristics, attracts more and more interest. In this study, we pay attention to vein pattern certification technology. We firstly take photos of a vein pattern of forefinger using a transmitted near-infrared light method with a near-infrared light camera. We then convert them to grayscale images, emphasize the vein part, and get normalized images. Finally, we propose some algorithms to compare two of these images. One of the most important problems that lower the accuracy of recognition is the difficulty of adjusting the location of the two images. We show some experimental results as a basis to solve this problem.