Machine learning models for crop yield prediction using satellite imagery allow accurate, reliable, and timely estimations of crop yield. This information is necessary to ensure the adequacy of a nation’s food supply as well as to aid policymakers and farmers in decision-making. In India, rice is one of the most important crops. The aim and objective of this research was to develop and evaluate machine learning models to estimate crop yield from satellite data. In the current study, the researchers have applied machine learning models to obtain rice crop yield in the rice-producing regions of India such as Andhra Pradesh, Uttar Pradesh, and West Bengal. Area, weather, irrigation, and temperature are the factors that impact crop yield. The area is computed using satellite imagery from the Google Dynamic World Earth data set that contains near real-time (NRT) satellite images. Weather, irrigation, NDVI, temperature, and season data were obtained from various sources. The collected data was analyzed using XGBoost Gradient, Random Forest, and Support Vector Regressor. The models were trained and tested. The results indicated XGBoost is the best model with a root mean square error of 80, 400 for rice, whereas R2 = 0.94 was for the same crop datasets. The findings of the study can facilitate agricultural decision-making for policymakers and farmers.