The analysis of land use and cover maps enables a comprehensive understanding of the territorial transformations associated with urban growth and their impacts on ecosystems. In this study, we employed the Google Earth Engine to pre-process and mosaic the Landsat-9 images from 2023 of the surrounding area of Guayaquil city, Ecuador. Furthermore, we compared the performance of three supervised classification algorithms: Random Forest, Support Vector Machine, and Artificial Neural Network using RStudio. Four land cover classes were defined: forest, crops, no vegetation (buildings, roads and bare soil), and water. Training samples were obtained through visual interpretation and cross-verified with MapBiomas. Validation of the results was conducted using the Kappa coefficient and overall accuracy, compared against reference maps from MapBiomas. The findings indicate that the Support Vector Machine algorithm achieved the highest accuracy (Kappa = 0.91; OA = 93%), slightly surpassing Random Forest (0.89; 92%) and Artificial Neural Network (0.86; 90%). These results confirm the robustness of the SVM algorithm relative to the other methods and highlight its potential for urban monitoring in tropical environments. The methodology employed integrates open satellite data, reproducible tools, and machine learning algorithms, thereby supporting the selection of the most suitable algorithm for the conditions of Ecuador's coastal cities and contributing to sustainable territorial planning aligned with Sustainable Development Goals 11 and 13.