This paper proposes a Forgetting Factor-based Multi-Innovation Extended Kalman Filter (FMIEKF) algorithm to address the challenges of accuracy and stability in multi-source external measurement data fusion. Traditional Extended Kalman Filter (EKF) algorithms often struggle to meet the requirements of high-precision data fusion in nonlinear systems due to their sensitivity to noise and limited fusion accuracy. To overcome these limitations, this study introduces an improved filtering algorithm that integrates multi-innovation identification theory with a forgetting factor mechanism. The FMIEKF algorithm employs a sliding window technique to fully utilize historical data while incorporating a forgetting factor to dynamically adjust the influence of past observations on current estimates. This approach effectively mitigates the interference of noise in historical data, thereby enhancing the accuracy and robustness of parameter estimation. In the implementation of the algorithm, the FMIEKF extends the innovation sequence of the traditional EKF through multi-innovation theory, increasing the efficiency of information utilization and further improving the convergence speed and estimation accuracy of the filter. Simulation results demonstrate that the FMIEKF algorithm significantly outperforms the conventional EKF in terms of mean square error (MSE) for both position and velocity filtering. Notably, the proposed algorithm maintains high stability and reliability under varying noise conditions, a critical advantage in dynamic environments. These characteristics highlight the potential of the FMIEKF algorithm for broad application in aerospace tracking and control systems, offering a novel and effective solution for multi-source data fusion in complex scenarios.