In the last decades, the interest in thin hydrocarbon reservoirs has grown progressively justifying the great effort spent on developing techniques for quantitative interpretation of thin beds seismic response (Widess, 1973; Kallweit and Wood, 1982). Neidell and Poggiagliolmi (1977) and Meckel and Nath (1977) identified seismic amplitude and apparent thickness as the key elements to describe complex waveforms in thin layers environments. They suggested using this information to predict net thickness wherever well data are available for calibration. This approach was further developed introducing a stratigraphic modelling procedure to systematically investigate the relationship between layers geometry-lithology and seismic response (Schramm, Dedman, and Lindsey, 1977). In the mid eighties, interactive interpretation using workstations allowed the quantitative analysis techniques to be extended to 3D seismic data. Brown et al. (1986) in their classic publication proposed using statistical tuning curves (derived by interactive cross plotting) and deterministic curves (by wavelet extraction) to remove the geometric effects of thin beds from seismic amplitudes. The detuned amplitudes were then input to map the net gas sand thickness. In the recent past, Neff (1990 and 1993) has implemented a workflow for reservoir characterization based on seismic and petrophysical modelling (incremental pay thickness modelling). This approach enable the mapping of gross pay, net pay, net porosity, and hydrocarbons in place for clastic and carbonatic reservoirs. Conceptually, this procedure is absolutely necessary whenever layer thickness makes the amplitude interpretation less intuitive and the definition of a reservoir quality indicator more difficult. In this article we have extended the analytical approach for thin beds evaluation to the three AVO classes (Rutherford and Williams, 1989). Since we were mainly interested in the methodological aspects of seismic thin beds characterization, we have chosen to follow an approach based on both forward and inverse wedge modelling, that combines rock physics and tuning analysis for a petrophysical interpretation of the tuning curves. For each AVO class we defined a reference model starting from deepwater real data. Then we investigated (1) which reservoir quality indicator (net to gross, porosity, net thickness, and porosity thickness) could be predicted from seismic amplitude and apparent thickness and be quantified after tuning charts calibration and (2) how reliable are two common practices of seismic lithology, namely reservoir quantification (i.e. prediction of specific petrophysical parameters from seismic derived elastic attributes) and classification (i.e. facies discrimination based on the inverted elastic attributes). Finally, we illustrate two examples of thin beds reservoirs quantification both in the development and appraisal phases.