<p>In the Laoshan Uplift of the South Yellow Sea, Cenozoic terrestrial clastic strata directly overlie Mesozoic-Paleozoic marine carbonate rocks, thus forming a strong impedance contrast layer that causes significant attenuation of seismic energy. The shallow-water environment further degrades the signal-to-multiple ratio (SMR) in middle-to-deep subsurface targets. To address the challenge of improving seismic imaging quality, this study presents an effective multiple suppression technique. It first analyzes the formation mechanisms and main categories of multiples within the shallow-water strong impedance contrast layer, indicates that primary reflections beneath this layer exhibit weak energy whereas multiples carry strong energy. Near-to-medium offset multiples mainly consist of seafloor multiples and surface-related multiples (SRM) associated with the strong impedance contrast layer. Moreover, medium-to-far offset multiples are primarily dominated by long-period multiples and internal multiples. In response, this study proposes a frequency-domain adaptive subtraction method, which uses a least-squares multiple prediction filter to optimize the prediction of seafloor multiples and SRM. By matching predicted multiples with input data within each time-frequency window, the filter refines the multiple model while preserving primary energy integrity, thereby effectively suppressing multiples generated by dipping formations and phase shifts. Field data applications of this integrated multiple suppression strategy significantly improves the SMR in the shallow-water strong impedance contrast layer of the Laoshan Uplift, enhancing the imaging quality of Mesozoic-Paleozoic targets, and clearly delineate the discontinuities between Sinian and Cambrian strata beneath the strong impedance contrast layer. These results contribute new understanding into the tectonic and sedimentary evolution of the Lower Paleozoic in the South Yellow Sea.</p>