The invention discloses a distribution transformer centralized intelligent detection method comprising the steps that multiple different terminals acquire the characteristic data of the transformer and then transmit the data to a master control computer; the master control computer summarizes the data and analyzes the characteristic data of the transformer by using a neural network algorithm withcombination of the historical data; and consistency verification is performed on the characteristic data and the data are added to a database after passing of verification. The beneficial effects of the distribution transformer centralized intelligent detection method and system are that as many as five transformer detection devices are integrated on one platform by using the advanced serial communication technology and five transformer parameters are detected in one step so that the detection program can be simplified and the detection efficiency can be enhanced; meanwhile, multiple detectiondata are analyzed by using the advanced algorithm with combination of the historical experience data so that the accuracy and the reliability of the transformer characteristic detection and analysisresult can be enhanced; besides, the system is enabled to have the self-improving function by continuous expansion of the historical database.