In recent years, the dynamic gravity measurement technology utilizing optical-gyro inertial navigation systems has advanced rapidly, owing to its high efficiency capabilities in acquiring Earths gravity field data. However, existing filtering methods exhibit limitations in addressing the frequency-domain aliasing of gravity signals and noise, which severely constrains the accuracy and resolution of measurements. To address these challenges, a novel joint noise reduction method integrating EMD (empirical mode decomposition) and WCF (wavenumber correlation filtering) was innovatively proposed. The raw gravity results were low-pass filtered initially. The filtered results then underwent EMD, and a threshold was applied to retain low-order intrinsic mode functions. Subsequently, correlation filtering was implemented on repeated line retention results to reconstruct signals and suppress noise. Simulation experiments demonstrate that the proposed method achieves a 44% improvement in signal-to-noise ratio and a 48% reduction in mean square error compared to traditional EMD processing. These results confirm the effectiveness of the EMD-WCF method in balancing measurement accuracy and resolution, offering a new strategy for high-frequency gravity signal denoising.