<p>Remote sensing satellites in low earth orbit are subject to various perturbation forces, frequently necessitating orbital maneuver operations. During these maneuvers, the introduction of maneuvering forces causes abrupt changes in the dynamics model, thereby degrading the accuracy of orbit determination based on traditional Extended Kalman Filter methods. To address the accuracy degradation caused by dynamics model inaccuracies during maneuvering, this paper utilizes an onboard accelerometer to measure the acceleration generated by non-conservative forces. Specifically, an Extended Kalman Filter incorporating the Savitzky–Golay Filter algorithm is proposed for the error calibration of accelerometer data, thereby enhancing the measurement accuracy and providing more precise data to support accelerometer-assisted orbit determination. Furthermore, to prevent filter divergence, a continuous maneuvering orbit determination method with a covariance matrix reset strategy is introduced. Experimental results show that in continuous maneuvering scenarios, the position root mean square error of the proposed method is 0.668&#xa0;m. This demonstrates that the proposed method can successfully resolve the filter divergence issue during continuous maneuvering, effectively improving the accuracy and robustness of orbit determination.</p>