Abstract: To address the challenges of high randomness in power source output and pronounced power oscillation in power grids, this study investigates a primary frequency regulation control method for generating units based on multi-source data fusion. By integrating heterogeneous measurement data from multiple sources, the system enables real-time dynamic and precise adjustment of frequency regulation capacity for each generation unit. The introduction of adaptive energy storage droop control ensures both safety and economic efficiency during frequency regulation. Through the implementation of a "fast-response slow-return" logic, the dynamic response process of frequency regulation power is optimized, effectively suppressing secondary frequency drops and power oscillations. Comparative experiments demonstrate that this method not only enhances frequency stability but also significantly mitigates power fluctuations, exhibiting superior control performance.