Neutron lifetime logging has been widely used, as the oilfield enters the mid and late development. Aiming at the existence of neutron lifetime logging curve depth error, contrast correction curve of two different stratigraphic information, and take advantage of the characteristics of logging curve sequential, for logging curve of hierarchical feature extraction and calibration. Logging curve correction method is presented based on the hierarchical feature extraction. Using the maximum position, spike number to match and categorize, and then using the maximum thickness ratio, and relative slope characteristics of the curve segment to judge similarity, and Finally realize the hierarchical feature extraction, these all based on the experience of the artificial vision correction. The experimental results show that the method can quickly and efficiently correct the logging curves.