Automatic digitization of well logs is a very relevant and important problem, because there are many paper well-logging diagrams in oil companies obtained before the computer age and though having no digital analogs. Such data need to be presented in the digital form for their computer analysis and further interpretation. Existing software for well logs processing digitizes well logs in semi-automatic mode, and this is not effective because of influence of the human factor and a large amount of information, which need to be processed. The authors have designed a program that implements automatic digitization of well logs with filled well passport containing information about border changes for every log and depth intervals. Testing of developed program revealed obtaining required accuracy withal increase of calculation speed in comparison with used software for semi-automatic digitization. At the moment identification of an artificial neural network for solving the problem of pattern recognition, which implements the problem of automatic digitization for diagrams headers and depth intervals, without filling well’s passport, is being held. It is planned further to use developed software for digitizing logs from paper and re-digitizing logs having been processed earlier in semiautomatic mode, i.e., microprobes readings. And this will allow increasing accuracy of geological model construction and hydrocarbon reserves estimation.