Although major care is usually taken during the steps of acquisition and processing regarding the quality of the final seismic data volumes, reservoir geophysicists still face difficulties related to remaining noise and irreducible indetermination, as observed when calibrating seismicto well data or interpreting seismic attributes for reservoir characterization purposes. Geostatistics offers alternative methods concerning the analysis of the seismic information, using a probabilistic approach based on the analysis of the spatial variability of the data. In this paper, the Factorial Kriging (FK) technique is presented and applied in three different aspects in the seismic reservoir characterization workflow: (i) analysis and decomposition of geometrical attributes to improve fracture mapping; (ii) seismic noise characterization to generate more realistic petro-elastic models in time-lapse feasibility schemes; and (iii) improved time-lapse interpretation strategies by FK decomposition. This filtering technique is suited for spatial analysis and it is shown that a great improvement of seismic data quality is achieved.