Rainfall variability remains a critical source of uncertainty in agricul tural systems, particularly in climate-vulnerable regions where inaccurate fore casts can result in crop yield losses. Existing rainfall prediction models generate deterministic outputs, lacking uncertainty quantification and interpretability for high-stakes agronomic decision making. To address these limitations, we propose an Explainable AI-Driven Probabilistic Rainfall Forecasting and Harvest Opti mization framework (XAI-PRFHO). The framework integrates a Temporal Con volutional Network–Bidirectional Long Short-Term Memory (TCN–BiLSTM) architecture with conformal prediction to generate calibrated prediction intervals with statistical coverage guarantees. Model interpretability is achieved through a dual-layer mechanism combining SHAP (SHapley Additive exPlanations) and instance-level Integrated Gradients, enabling seasonal interpretation of forecast drivers. The model is trained and validated using meteorological and agro-envi ronmental datasets, including climate data, reanalysis products, and ground sta tion observations, to evaluate its robustness and generalisability across agroeco logical conditions. The probabilistic outputs are incorporated into an XGBoost based crop phenology model and optimised using the Non-dominated Sorting Genetic Algorithm III (NSGA-III) to derive Pareto-optimal harvesting windows that minimise rainfall-induced crop-loss risk and maximise expected yield. Ex perimental evaluation demonstrates that XAI-PRFHO achieves a 7-day-ahead Mean Absolute Error (MAE) of 3.41 mm/day, outperforming evaluated baselines by up to 33.4%, while maintaining an empirical prediction-interval coverage of 94.7% at the 95% nominal coverage level. Harvest scheduling guided by the framework reduced model-estimated weather-induced crop losses by 31.7% rel ative to conventional farming heuristics, corresponding to a projected seasonal yield gain of 11%. The framework provides an integrated and interpretable deci sion-support approach for precision agriculture for climate-resilient farming sys tems.