The image classification procedure to identify remote sensing signatures from a particular geographical region can be performed with an identification model that has the ability to use large datasets to reach an accurate result. This novel methodology is referred to as the Statistical Enhanced Classification algorithm, which has been developed to employ multispectral images based in the statistical supervised learning theory and can be used for applications in environmental monitoring and analysis. This paper presents the performance study of the proposed methodology using both, multispectral synthetic images and multispectral remote sensing images. The obtained results are accurate due to the use of several spectral bands, the use of statistics such as mean and standard deviation for the training classes and for the pixel neighborhood, which provides more robust information, and the decision-making rule that has the ability to decide if the pixel is not belonging to a predefined class, which leads to an accurate decision model.