Abstract: In this article, taking the oil-immersed power transformers as the research object, the design scheme for transformer automation online detection system is proposed to address the problems of traditional transformer monitoring and diagnosis systems such as single detection parameters, poor real-time performance, and low fault diagnosis accuracy, etc. On the basis of the transformer oil chromatogram monitoring function, the system adds partial discharge and temperature detection modules to achieve comprehensive status monitoring and fault diagnosis of transformers. At the same time, the offline gas parameter measurement scheme is optimized, and the communication network architecture is used to achieve real-time transmission of multiple parameters online, effectively improving the intelligence level of the transformer detection system.