Understanding the spatial distribution of Land Use and Land Cover is crucial for promoting sustainability. This research leverages remote sensing to compare six approaches for mapping rice paddies and other Land Use and Land Cover types within a study area in the Cuenca Laguna Merín region of Uruguay. The study’s objectives include evaluating the performance of six classifications using Sentinel-2 and Sentinel-2/Sentinel-1, identifying the benefits of integrating microwave and optical data, assessing the effectiveness of short-term optical time series, and determining feature importance. The materials include quoted imagery, samples of classes, Random Forest, Google Earth Engine, and Python’s Scikit-learn GridSearchCV. The methods involve hyperparameter tuning, model development, supervised classification, and accuracy assessment. Maps from data fusion perform better than those based solely on optical classifications. The main advantages of this fusion are better identification of rice paddies and enhanced ability to distinguish between herbaceous and other summer crops. The importance of features reveals that near-infrared, short-wave infrared and various indices are more significant than microwave data; however, the latter plays a crucial role in achieving more accurate outcomes. Creating Deep Learning models and extending similar research to all the Cuenca Laguna Merín and other Uruguayan rice regions is recommended.