AVO/AVA (Amplitude Versus Offset/Angle) analysis has been used to determine the physical characteristics o f shallow marine sedi ments. The parameter that can be constrained, are P-and S-wave velocity, bulk density and acoustic attenuation. Shear modulus and other geotechnical properties such as grain size can be inferred from these parameters. The theory o f AVO investigations is given by the Zoeppritz equations (Zoeppritz, 1919), which allow the com putation o f the reflection and transmission coefficients as a function o f angle o f incidence. AVO has been w idely used in the petroleum industry, usually for frequencies o f 20-100 Hz. High frequency shallow sediment AVO studies require special analysis including careful geometry and source and receiver directivity corrections. In the past, marine sediments have been m odeled as elastic materi als. However, viscoelastic models with absorption are more realis tic. At large angle o f incidence, AVA-fimctions derived from vis coelastic models differ significantly from those with purely elastic properties. The influence o f S-wave velocity on the reflection coef ficient is rather small (especially for the low S-wave velocities encountered at the seafloor). On the other hand, P-wave velocity and density show a considerably stronger effect (Fig. 1). Thus, it will be difficult, or nearly impossible, to extract the S-wave param eter from AVA trends, especially when the signal-to-noise ratio is low.In order to measure the reflection coefficient in a seismogram, the peak amplitudes o f the direct wave and the seafloor reflection in a CMP (Common Mid Point) gather are determined and corrected for spherical divergence, and source and receiver directivity (Fig. 2). At CMP locations showing different AVA characteristics, the sedi ment parameters P-and S-wave velocity, density and absorption are determined by using a Sequential Quadratic Programming (SQP) inversion technique. The use o f the viscoelastic model within the inversion gives better results in fitting the data than the elastic model without absorption.The introduction o f constraints, e.g. empirical relationships given by core investigations, can improve and stabilize the inversion results.In Fig. 2, a soft mud layer, with a typical low P-wave velocity close to w ater velocity, forms the seafloor at CM P-Bin 570 whereas at CMP-Bin 4000 the seafloor contains boulder clay, a mixture of sand and gravel with a higher P-wave velocity and density. In both cases the inversion gives reasonable results for the first three sediment parameters. Note that the value o f the S-wave veloc ities have an uncertainty o f more than 100% due to the insensitivi ty o f the reflection coefficient to this parameter.The investigation o f deeper sediment layers by AVA requires a good estimate of the absorption (described by the Q-factor) within the first sediment layer.The influence of Q on the reflection coefficient is enhanced in the wide-angle domain. However, our measurements were restricted to a maximum angle of incidence of 60°, which...