Abstract: The spatiotemporal evolution of the Global Ionospheric Total Electron Content (TEC) directly determines the positioning accuracy of Global Navigation Satellite Systems (GNSS) and the reliability of communication systems. Achieving precise TEC prediction remains a key challenge in space weather research. However, existing ionospheric forecasting models predominantly rely on traditional solar activity indices and geomagnetic indices as drivers. These traditional indices often suffer from low temporal resolution and infrequent updates, thereby limiting the forecasting performance of data-driven models under extreme space weather conditions. To address these issues, this study proposes an Index Group based on the Spectral Whitening Method (SWM). By innovatively introducing a high-latitude disturbance index, the system provides a full-chain characterization of solar-terrestrial energy coupling. A two-stage deep learning framework, integrating 3D electron density reconstruction (DNN) and spatiotemporal prediction (3D Swin Transformer), was constructed for validation. Experimental results demonstrate that SWM indices maintain high correlation with traditional indices while significantly accelerating model convergence through their high-frequency physical features. The SWM-driven model significantly outperforms the IRI-2020 empirical model and effectively suppresses error accumulation in long-term forecasting. The study confirms that the SWM system, with its superior real-time update capability and sensitivity to disturbances, offers a novel technical pathway for operational global space weather forecasting.