We apply instantaneous seismic attributes to a stacked P-wave reflected seismic section in the Tenerife field located in the Middle Magdalena Valley Basin (MMVB) in Colombia to estimate effective porosity f, water saturation Sw and volume of clay Vclay at seismic scale. The well-logs and the seismic attributes associated to the seismic trace closer to one of the available wells is the information used to train some multi-layered Artificial Neural Networks (ANN). In order to chose the best input combination of seismic attributes to train an ANN to estimate these petrophysical parameters, we perform data analysis using a mathematically non-parametric nonlinear smooth modeling tool: the Gamma test. Once the ANN is trained it is applied to predict these parameters along the seismic line. This is a significant result that shows for the first time a petrophysical characterization of this field at seismic scale. From the continuous estimations of Vclay we distinguish two facies: sands and shales, these estimations confirm the production of the Mugrosa C-Arenas. This classification into two facies allows us to select only the sands or shales and plot the porosity or the Sw for each one of the facies separately. These estimations are particularly useful because we can distinguish an interesting lateral variation of Sw and hence a variation of the hidrocarbon saturation (Sh) in the paying zones since Sw+Sh=1.