We present a true-azimuth 3D free-surface multiple attenuation workflow tailored to handle complex seafloor topography in shallow-water environments. The workflow is based on a modified approach proposed by Moore and Bisley (2006), where the key change lies in the application of a multidimensional convolution framework to predict all free-surface multiples. The initial step consists of attenuation of all multiples with a bounce in the water layer using a hybrid wavefield extrapolation based on a Green’s function of the seafloor. The second step attenuates remaining free-surface multiples. We show one example from a North Sea multimeasurement towed-streamer survey that directly benefits from dense receiver coverage to circumvent limitations commonly found on conventional towed-streamer data due to sparse sampling of shallow-water events, while enabling the computation and application of multidimensional adaptive matching filters. Results show that we are able to accurately predict and attenuate all free-surface multiples.