This paper presents the application of a novel variant of the Kalman Filter algorithm, known as the Local Ensemble Transform Kalman Filter (LET-KF), for estimating and mitigating power system harmonics in the presence of random noise. The proposed algorithm improves efficiency and accuracy compared to traditional Kalman Filter methods by reducing computational complexity and storage requirements. Results demonstrate that LET-KF provides superior performance in estimating harmonic parameters across various noise levels and dynamic signal conditions.