Abstract: The support vector machine arithmetic, based on statistical theory, is presented in study of recognition of oil, gas and water zones from geologic logging. The parameters from gasometry logging and geochemistry logging and relations among induced parameters are studied, thus developing a model with help of LibSV M software, by which a novel method for recognizing oil gas/ water zone from geologic logging is proposed. The real application of this method to an oilfeld in Xinjiang shows that the support vector machine arithmetic is simple,reliable, more practical and higher accurate for use.

Key words: logging, method, recognition, oil-gas zone