The remaining useful life (RUL) of complex systems and critical components is a key metric for evaluating their reliability and maintainability. Accurate RUL prediction enhances maintenance efficiency, reduces costs, and improves overall operational safety. Traditional physics-based models for RUL prediction typically rely on fixed parameters, which limit their adaptability and real-time effectiveness. Additionally, parameter estimation in these models is often based on complete degradation data. While this approach captures the overall degradation trend, it fails to account for fluctuations throughout the degradation process. To address this limitation, [14] introduced an RUL prediction method that integrates particle filtering (PF) with degradation modeling, allowing for dynamic updates of model parameters. However, this method does not account for the issue of particle weight degradation during the resampling process in PF, which can compromise prediction accuracy. To mitigate this issue, this paper proposes replacing the standard PF algorithm with a particle swarm optimized particle filtering (PSO-PF) algorithm, aiming to enhance prediction accuracy. The effectiveness of the proposed approach is validated using a publicly available lithium battery dataset.