This paper introduces an efficient approach, the dynamic coefficient polynomial model, which emulates crop growth dynamics using NDVI. This model, a significant improvement over traditional models like the NDVI mean and static polynomial models, is designed to be adaptable over time and incorporates spatial variables to account for the diverse growth conditions experienced in different regions. Consequently, the model’s responses and adaptations are influenced by the specific crop growth dynamics observed within these spatial dimensions, adding a new dimension to crop growth forecasting. Our results show that the proposed model achieves a higher accuracy than the other machine learning models, which is about 90.3%.