Image registration is an important preprocessing procedure for remote sensing image applications, such as geometric correction, change detection, and image fusion. Since it is a time-consuming and labor-intensive task to correctly register the remote sensing image, this paper proposes a fully automatic and robust approach for the remote sensing image registration. First, the image pyramid of working and reference images are constructed for coarse to fine matching processing. Second, the feature points can be automatically extracted from the reference image, and the matching point can be searched on the working image. Third, in order to improve the accuracy of registration, the robust estimation serves as an important tool in preserving the correctly matched points. Three sets of satellite images, which include multi-sensor, multi-temporal and multi-spectrum images, are used to test the proposed approach. Results show that the approach is capable of automatically registering the working image to the reference image with great precision.