Quantitative petrophysical characterization is one of the principal tasks of a reservoir analyst and is generally affected by the methods used. For example, different theoretical and empirical formulas are in most cases restricted to the specific areas where they were developed. Prediction of continuous petrophysical parameters is often time consuming and complicated because of geological variability such as facies changes due to sedimentary and structural changes. In this work we propose a neural network approach which is used to predict quantitative petrophysical parameters from wireline logs of cored intervals. We then apply the knowledge learned during training to uncored intervals or other holes. Data from the Ocean Drilling Program and from two South American oilfield holes are used to test this technique. The results show a good match between the neural network-derived petrophysical parameter logs and the actual core measurements. Problematic petrophysical measurements can be identified by a mismatch between the responses.