Abstract: To investigate the causes of deviation between numerical simulations and experimental measurements of the flow and acoustic fields from underwater cavities, this study analyzes the influence of deviations in shear oscillation frequency and amplitude. Firstly, numerical simulations overestimate the shear oscillation frequency due to the neglect of structural damping in the cavity model. Secondly, the non-closed-loop nature of acoustic feedback at the cavity orifice suppresses the formation of higher-order shear oscillation peaks in experiments. Thirdly, the acoustic cutoff frequency effect in the test environment leads to deviations between the computed and measured self-noise inside the cavity. Finally, a method for adjusting numerical parameters is proposed to mitigate these deviations. Experiments conducted in a circulating water channel to measure unsteady pressure fluctuations demonstrate that optimizing mesh parameters reduces the deviations in shear oscillation peak levels to within 3 dB. This study provides a theoretical foundation for accurate prediction, deviation correction, and line-spectrum noise suppression in cavity flow–acoustic interactions.