A new approach is presented here for an automatized stratigraphic subdivision of well logs into high-frequency sequences. This method allows the gap to be bridged in data processing between primary, uninterpreted well data and more complex stratigraphic models. The principal advantage oi this simple, but soundly based approach is its robustness and rapidity in routine stratigraphic studies where extensive data bases are analysed. This paper presents the basic procedures oi an automated approach to genetic sequence stratigraphy, which permits the full exploitation of available well logs and cores (or outcrop sections). The first and most critical step is to transform log data, by statistical cross-plot analysis, into geologically meaningful electrofacies. Use of the most plausible facies model then helps to re-define the different electrofacies in terms of paleo water-depths. The well log thus provides a paleobathymetric curve with successive maxima and minima (i.e. definitions of maximum flooding and transgressive horizons). Sudden decreases in paleo water-depth are marked as horizons of maximum base level fall and, therefore, as potential sequence boundaries. ii no such sudden events are recorded, then the minimal water depth is seen as synonymous with the sequence boundary. Procedures such as the automatic identification of larger correlatable cycles, and the construction of Fischer diagrams (corrected for paleo water-depths) allow to establish reliable well correlations. Finally, classical Fourier analysis (using the accommodation curves oi various wells) helps in establishing a synthetic eustatic curve.