One of the common challenges in image satellite classification prediction models is obtaining high accuracy, and from this perspective, the primary objective of this paper is to implement the Stacked Learning Model (SLM) technique with the aim of enhancing the accuracy of image classification. Academic research and practical applications have widely used the suggested model, which has made significant strides in the field of image classification. By using a SLM, the model combined the results of three base models to improve the predictive ability and accuracy of image classification. This approach had been applied in a series of sequential steps designed to improve the performance of individual models and their integration. The proposed model uses the EuroSat dataset, which includes images. These images are of high resolution and cover various land use and land cover classes across Europe. They have been evenly distributed across six classes: residential, permanent crop, pasture, industrial, highway, and forest. This data has been augmented to reduce the problem of overfitting and improve the model’s ability to generalize. The model achieved a classification accuracy of 92.35% after training, confirming its robustness in image classification applications, particularly in areas where accuracy is crucial. Various comparisons, even with state-of-the-art models, will be provided.