Image panorama is a process of combining two or more images to form one single image. Image stitching is a process of creating image panorama from a set of images with overlapped fields. Image stitching faces many challenges such as images corrupted by noise, indexing a large number of images, high image resolution, and presence of parallax and scene motion. The problem in eliminating visible seam is also another challenge. To obtain a wide seamless panorama, this paper implemented a feature-based automatic image stitching algorithm. The image stitching model consists of five stages: image registration, features detection, feature matching, Homography estimation, and image blending. A method of creating a seamless image panorama was introduced where the scale-invariant features transform (SIFT) was used for image feature extraction, the K-nearest neighbor algorithm for feature matching, the Random sample consensus (RANSAC) for image warping calculating homography and the weighted matrix was intended for image blending. Three (3) sets of source images were used to test for stitching demo. The effectiveness of the stitching method was defined by comparing the resulting panorama to a predicted reference image using the percent image similarity defined as the proportion between the matches found and keypoints. For all of the testing sets, the resulting percent similarity with respect to the reference image ranges from 12-18%. This is due to the fact that high-quality images (i.e. high number of pixels) have thousands of features -hence, thousands of keypoints while low-quality images may have only a few hundreds.