The State of Charge (SoC) of an electric vehicle plays a crucial role in its performance. SOC can be used to accurately formulate an electric vehicle charge control strategy, which can improve the vehicle’s endurance, reliability, and cost. This paper introduces a new feature parameter as an improvement to the Adaptive Forgetting Factor Recursive Least Squares (AFFRLS) algorithm. The new parameter is designed to capture the dynamics of innovation, thus optimizing the battery model parameter estimation process, and improving its accuracy. Moreover, a function with adjustable parameters was developed with the objective of dynamically determining the optimal value of the forgetting factor. In this paper, the improved AFFRLS algorithm is referred to as the IAFFRLS algorithm. Additionally, we enhance the Adaptive Extended Kalman Filter (AEKF) algorithm by introducing an innovative threshold detection mechanism, thereby optimizing its estimation process. Specifically, the Sage-Husa estimator is employed to adjust the process noise covariance and the measurement noise covariance when the innovation falls below a specified threshold. Conversely, when the innovation exceeds the defined threshold, the algorithm maintains the current parameter state to avoid unnecessary adjustments, effectively minimizing computational burden and the risk of error accumulation during the estimation process. Similarly, the improved AEKF algorithm is hereafter referred to as the IAEKF algorithm. Furthermore, a process for lithium battery SOC estimation using the IAFFRLS-IAEKF joint algorithm is described. To conclude, the results of a series of simulations and experiments demonstrate that the IAFFRLS parameter identification accuracy is higher than the AFFRLS algorithm, and the IAFFRLS-IAEKF joint algorithm outperforms the IAFFRLS-AEKF joint algorithm in terms of accuracy. The effectiveness of the algorithm proposed in this paper has been validated.