Abstract: Currently in the field of high voltage equipment, intelligent configuration guidance is not an effective method of oil-immersed power transformers in this example, a power transformer to explore the theory of intelligent way configuration. For 7 kinds of the power transformer state classification, including normal operation, failure, winding core failure, casing failure, the failure of the cooling system, tap-changer failure, failure, such as insulating medium according to the statistics of the calendar year power transformer reliability data, in this state of 7 kinds of the construction of a transformer of markov process, to calculate the stationary state markov probability. The observation matrix is constructed to calculate the probability that 7 kinds of transformer faults can be detected by the monitoring sensor. The greater the probability, the more important the monitoring sensor is, and the online monitoring parameters of the intelligent transformer are sorted according to the importance of the sensor. The results show that this paper provides a theoretical method with strong operability for the on-line monitoring equipment configuration of intelligent transformer.