Climate change creates far-reaching environmental, economic, and social impacts across the globe, making reliable and timely prediction tools essential for effective mitigation and adaptation efforts. This research introduces a machine-learning-based climate prediction system that utilizes both historical records and realtime environmental data to project climate patterns. The framework combines data preprocessing, feature selection, and predictive modeling to uncover complex nonlinear interactions among major climatic factors, including temperature, precipitation, carbon emissions, and atmospheric variables. Several machine learning models are assessed to identify the most suitable approach for long-term climate forecasting. Results from the experiments indicate that the proposed system delivers strong accuracy and consistency when compared with conventional statistical methods. Overall, the findings demonstrate how artificial intelligence can strengthen climate prediction, inform environmental policy decisions, and support sustainable development goals.