Disclosed is an online monitoring and diagnosis model self-adaption algorithm used in a transformer substation. The online monitoring and diagnosis model self-adaption algorithm is used for self adaption and optimization of an online monitoring and diagnosis model of the transformer substation. The online monitoring and diagnosis model self-adaption algorithm comprises an online monitoring and diagnosis model dynamic link library, an XML configuration file and a self-adaption algorithm. According to the technical scheme, diagnosis models which are used for online monitoring of the transformer substation are complied into a form of a dynamic link library, then are directly copied on a memory space of online monitoring equipment or diagnosis equipment, then the corresponding XML configuration file and the self-adaption algorithm are matched up to achieve the purpose of judging whether the diagnosis models of various kinds are utilized or not, the priority level and the demand of model interfaces, and thus the voluntary matching and self adaption of the online monitoring diagnosis models of various kinds are achieved. According to the online monitoring and diagnosis model self-adaption algorithm used in the transformer substation, the self-adaption calling of the diagnosis models of various kinds and categories can be achieved through the algorithm, and the difficulty that the diagnosis models of various kinds cannot conduct self adaption and can only be used alone is solved; meanwhile, the algorithm is simple and easy to use, convenient to operate and high in using efficiency.