Panchromatic and multi-spectral images produced by the remote sensing satellites are fused together to provide a multi-spectral image with a high spatial resolution at the same time. The spectral quality of the fused images is very important because the quality of a large number of remote sensing products depends on it. Due to the importance of the spectral quality of the fused images, its evaluation is also important. This paper presents an object-based strategy for evaluating the spectral quality of fused images, aiming to overcome the limitations of the current pixel-based method. This type of assessment is conducted by focusing on homogeneous objects with similar spectral and textural behaviors. In the implementation phase of the article, after determining an optimal spectral metric, the proposed object-based strategy is applied to five datasets from four different satellite sensors types, and the spectral behavior of the fusion methods has been studied in several image classes. The results indicate that the spectral behavior of the fusion methods does not follow a deterministic rule. Finally, statistical analyses were used to determine the best fusion algorithms in each class, and a list of superior algorithms in different classes was provided to researchers in the field of remote sensing and image fusion. This approach will help the scientific community to take a realistic vision at choosing the fusion algorithm appropriate to their satellite imagery.