The ability to quickly and accurately estimate the state of charge (SOC) of lithium batteries is a key function of the battery management system (BMS). To enhance the accuracy of SOC estimation for lithium batteries, we propose a method that combines the dynamic factor recursive least squares (DFFRLS) algorithm and the strong tracking H-infinity filtering (STF-HIF) algorithm. To address the issue of fixed forgetting factors in recursive least squares (RLS) that struggle to maintain both fast convergence and stability in battery parameter identification, we introduce dynamic forgetting factors. This approach adjusts the forgetting factor based on the residuals between the model’s estimated and actual values. To improve the H-infinity filtering (HIF) algorithm’s poor performance in tracking sudden state changes, we propose a combined STF-HIF algorithm, integrating HIF with strong tracking filtering (STF). Simulation experiments indicate that, compared to the HIF algorithm, the STF-HIF algorithm achieves a maximum absolute SOC estimation error (MaxAE) of 0.69%, 0.72%, and 1.22%, with mean absolute errors (MAE) of 0.27%, 0.25%, and 0.38%, and root mean square errors (RMSE) of 0.33%, 0.30%, and 0.46% under dynamic stress testing (DST), federal urban driving schedules (FUDS), and Beijing dynamic stress testing (BJDST) conditions, respectively.

Reliable battery state awareness is essential for energy management and power allocation in hybrid electric ships. However, battery management systems are increasingly exposed to False Data Injection Attacks (FDIAs) in intelligent connected environments, which can distort State of Charge (SOC) estimation and compromise the operational reliability of shipboard power systems. To address this challenge, this paper proposes a closed-loop “Modeling-Detection-Defense” framework for secure SOC estimation in marine cyber-physical energy systems. First, a stealthy FDIA model is developed based on battery dynamics and physical consistency constraints. Second, a hybrid detection method combining unsupervised and supervised learning is proposed to identify attacks. Finally, a long short-term memory network is employed to reconstruct compromised measurements and provide reliable SOC information for continuous energy management. Experimental results demonstrate that the proposed framework mitigates SOC estimation deviations caused by FDIAs. In addition, it effectively reduces power allocation errors and energy losses, thereby improving the cyber-resilience, operational reliability, and energy efficiency of hybrid ship power systems.