<p>The constant-offset seismic reflection method has been extensively utilized in geotechnical investigations due to its high acquisition efficiency and streamlined data processing workflows. Driven by the accelerating development of urban underground space, there is a growing demand to refine its application in complex, noise-intensive urban environments. To systematically elucidate the mechanisms by which urban settings influence imaging performance, this study established several representative numerical models, including a three-layer stratified medium, a model with a high-velocity overburden, and a model featuring underground utility conduits. Full-wavefield forward simulations were performed to quantify the sensitivity of imaging results to varied source wavelet frequencies and offset distances. The results demonstrate that source frequency dictates the critical trade-off between resolution and penetration depth: while high-frequency wavelets enhance the fidelity of shallow structures, low-frequency components are essential for identifying deeper targets. Furthermore, the offset distance primarily governs the signal-to-noise ratio (SNR) and deep-wavefield coverage; specifically, smaller offsets optimize near-surface resolution, whereas larger offsets are indispensable for intermediate-depth investigations. Notably, although surface high-velocity layers markedly degrade data quality, such effects can be mitigated through strategic parameter coupling. Overall, this research confirms the robustness and stability of constant-offset seismic reflection in urban exploration, providing a rigorous theoretical framework and optimized parameter guidelines for detecting subsurface anomalies and conducting municipal site investigations.</p>