The Cerrado is a highly diversified ecosystem and provides habitat for many species, however, it has undergoing marked degradation in recent decades due to the expansion of agricultural commodity production. This scenario reinforces the need for continuous monitoring of land use and land cover (LULC) changes, whether with a focus on environmentally sustainable agricultural production or market understanding. Recently, machine learning algorithms have become a promising and innovative approach to remote sensing data processing. Thus, this study aimed to evaluate the potential of the Random Forest image classification algorithm for LULC mapping and classification in the Brazilian Cerrado. The selected study area was the municipalities of Natividade, Chapada da Natividade, and São Valério da Natividade, located in the state of Tocantins. The basic materials of this study were the digital elevation model produced by the Shuttle Radar Topography Mission (SRTM), the night light images obtained by the Visible Infrared Imaging Radiometer Suite (VIIRS) sensor onboard the Suomi National Polar-Orbiting Partnership (Suomi NPP) and NOAA-20 satellites, and the Landsat 8 Operational Land Imager satellite (OLI) multispectral images acquired from May to October 2013. All analyzes were performed on the Google Earth Engine platform that allows cloud computing. A cube of images was generated containing 38 layers that were classified by the Random Forest algorithm, with 500 decision trees. For the algorithm training, random points from each LULC class mapped by the TerraClass Cerrado 2013 project were used. Considering the TerraClass Cerrado 2013 mapping as the ground truth, a Kappa index of 0.64 was obtained. There was a significant overestimation of annual cropland and urban areas. The proposed methodology presented a good potential for less expensive and less time demanding LULC mapping of the Cerrado.