Abstract: Oceanic internal flow disturbances and attitude adjustments of underwater vehicles can easily induce non-steady states in the Doppler Velocity Log (DVL), leading to degradation in its velocity measurement accuracy. Currently, methods such as attitude cosine matrix compensation and inertial navigation-assisted dynamic correction are employed to address this issue; however, problems including insufficient modeling fidelity and poor generalizability remain. To overcome these limitations, this paper establishes a seafloor echo model under non-steady conditions—incorporating the effects of echo pulse width, carrier velocity, and attitude on transducer signal transmission and reception. Based on this model, a simplified analytical expression for the echo Doppler shift under carrier non-steady motion is derived, and variations in frequency shift estimation parameters across different non-steady scenarios are systematically analyzed. A comprehensive correction method, grounded in frequency shift estimation parameters, is then proposed. Simulation and experimental data analyses demonstrate that the proposed echo model accurately captures the time–frequency characteristics of the bottom echo signal. Moreover, after correction using the proposed method, the cumulative comprehensive accuracy of DVL velocity measurement improves by more than 1.2%.