The global transition towards renewable energy sources and the increasing adoption of precision agriculture have intensified the need for accurate and real-time solar radiation forecasting systems. Conventional meteorological models often lack the temporal resolution and contextual relevance required for precise solar irradiance estimation. This deficiency can result in inefficient energy harvesting and poor decision-making in agriculture, especially in regions where solar power forms a critical component of the energy and farming infrastructure. This research proposes the development of an AI-powered Solar Radiation and Weather Prediction System that addresses the limitations of traditional forecasting methodologies. The system integrates heterogeneous data sources—including NASA's Prediction Of Worldwide Energy Resources (POWER) and OpenWeatherMap—to construct a robust environment for model training and validation. Employing advanced machine learning and deep learning techniques, specifically Extreme Gradient Boosting (XGBoost), Long Short-Term Memory networks (LSTM), and Transformer-based architectures, the system delivers high-resolution solar radiation forecasts with enhanced accuracy. Furthermore, the solution features a cross-platform user interface accessible via web and mobile devices. This interface provides users with real-time solar and weather predictions, timely alerts for hazardous conditions such as extreme UV exposure, and actionable agricultural recommendations, such as optimal irrigation scheduling. By offering precise, localized insights, the proposed system aims to improve the efficiency of solar energy utilization, support data-driven agricultural practices, and contribute to climate-resilient infrastructure planning.