The rapid growth in frequency and intensity of climate change events over past few decades present significant long-term challenges to agriculture such as crop yield, resource conservation and sustainability. This chapter investigates the role of machine learning (ML) models and hyper-parameter tuning optimization to improve climate resilience in agriculture. We extensively reviewed existing literature on climate resilience in agriculture in the view of emerging ML techniques. The methodology includes gathering meteorological and agricultural data to prepare dataset as foundation for innovation. The suitable machine learning models are selected, followed by hyper-parameter tuning approaches such as random search, grid search, and Bayesian optimisation. The results revealed that hyper-parameter tuning enhances the accuracy of ML models, leading to more accurate and reliable forecasts of climate impacts on agricultural yields. We compare the performance of tuned and untuned models across various evaluation metrics, showcasing the significant improvements achieved through detailed tuning. Our discussion further explores the practical implications of these findings, demonstrating how optimized ML models can guide adaptive agricultural practices and bolster resilience to climate variability.