The key of image matching is to establish the correspondence between two images. This can be achieved by using the constraint of epipolar line. The epipolar geometric relationship between two images can be described by the fundamental matrix. Because of the noise disturbulance and correspondence outliers, it is difficult to estimate the fundamental matrix. In this paper, a method based on the RANSAC algorithm for fundamental matrix estimation after coordinate normalization is put forward. Thus, the estimation error due to the mismatching points is eliminated. The result shows that the above method has a higher matching precision and can also be used to detect the objects in infrared images.