Predicting crop yield, which plays a pivotal role in crop market planning, insurance strategies, and efficient harvest management, presents a significant challenge attributable to the intricate interplay between various environmental and agricultural management factors. This study capitalizes on recent advancements in satellite technology and machine learning techniques to construct a robust prediction model tailored explicitly for the agriculture-intensive Hindon basin located in India. The method used in this study capitalises on the strengths of the Convolutional Neural Network (CNN) and Long-Short-Term Memory (LSTM) models, renowned for their expertise in capturing intricate spatial features and uncovering a variety of phenological traits crucial for accurate crop yield prediction. The model development phase involved training on a diverse set of variables encompassing crop growth indicators, environmental parameters such as MODIS Land Surface Temperature (LST) data and MODIS Surface Reflectance (SR) data, and historical crop yield records sourced from the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT). An analysis of the outcomes derived from the model indicates that the integrated CNN-LSTM framework exhibits superior performance compared to utilizing either the CNN or LSTM models in isolation. This advanced method holds great promise in enhancing the accuracy of crop yield forecasts, thereby empowering farmers to make informed decisions about the selection and optimal timing for growing specific crops.