Fusion of images with different spatial resolution can improve visualization of the images involved. This study tries to show that the fusion of the images from the same sensor system can also improve classification accuracy of the images. Four image fusion algorithms have been employed in the study of data fusion and classification of Landsat 7 ETM imagery, taking southeastern part of Fuzhou City as the study area. These are the Smoothing Filter-Based Intensity Modulation (SFIM), Modified Brovey (MB) Transform, Multiplication (MLT) Transform, and High-Pass Filter (HPF) Transform. The effectiveness of the four fusion algorithms has been evaluated based on spectral fidelity, high spatial frequency information gain, and classification accuracy. The study reveals that the SFIM transform is the best method in retaining spectral information of original image, which does not cause spectral distortion, and achieving the highest classification accuracy. MB-fused image has highest spatial frequency information gain but significantly loses spectral properties of the original image. The study shows all four fusion algorithms used can significantly improve the classification accuracy of the fused imagery. Therefore, fused images from the same sensor system can be used for improving not only visual interpretation but also classification accuracy due to free of the seasonal difference, various solar illumination and other environmental condition differences, and co-registration errors, which are common to the fusion using images from different sensor systems.