Crop monitoring and management of crops call for accurate crop categorization and precise yield assessment which is always critical. The conventional methodology of yield estimation such as the use of samples and ground surveys is time-consuming and may not be accurate spatially. In light of this gap, this study pulls together Google Earth Engine (GEE) to enhance these procedures through remote sensing and geospatial approaches. This study will therefore involve key crops, namely, turmeric, banana, and betel leaf using a methodology that entails preprocessing of satellite imagery, training classifiers such as random forest and smile cart using machine learning, and generation of yield maps based on the classified land cover classes. Using indexes and classifiers, the model produces maps with high accuracy of crop identification and the yields are estimated in sentiment from hectares to acre and incorporated to the classified maps which are further used for crop yield calculation to classify areas as low, medium, and high yield regions. The random forest model has shown 99.62% accuracy. This approach shows that these technologies have potential in helping on decision-making in agriculture and thus can be useful to the stakeholders.