The invention discloses a GRA and SVM based prediction method of power transmission and transformation project cost. At present, the relation between the power transmission and transformation projectcost and influential factors is hard to describe linearly. According to the prediction method, GRA is used to calculate the degrees in which the different factors influence the power transmission andtransformation project cost, the influential factor which changes with the power transmission and transformation project cost highly synchronously is screened according to the grey relational degree,a screened main influential factor serves as input of the SVM, a predicted value of the power transmission and transformation project cost serves as output of the SVM, the property that the SVM approaches the nonlinear continuous function relation infinitely is used, and the SVM serves as a prediction model of the power transmission and transformation project cost. The prediction method can be used to estimate the power transmission and transformation project cost accurately and reasonably, so that a power grid enterprise can improve the project cost control level and capability.