The invention provides an electric power metering equipment fault risk prediction method based on GBDT. The method comprises the steps that a fault risk prediction model and a feature set corresponding to a to-be-predicted electric power metering equipment sample are determined according to obtained basic data of the to-be-predicted electric power metering equipment sample, and the feature set comprises a fault class F, a fault class f, idle time T, in-operation time R and a property unit corresponding to the electric power metering equipment sample; taking the feature set as the input of a feature set fault risk prediction model, and determining the predicted life of the electric power metering equipment sample to be predicted, the predicted life being the cumulative use time corresponding to the specified electric power metering equipment fault probability; or the fault probability of the electric power metering equipment corresponding to the specified cumulative use time is determined. According to the prediction method provided by the invention, the predicted service life of the electric power metering equipment can be accurately determined, so that the operation state of the electric power metering equipment can be effectively monitored.