The invention discloses an electric energy meter failure rate assessment method. The method comprises the following steps of acquiring the failure data of an electric energy meter; selecting a corresponding B spline primary function order according to the failure data characteristics and the model complexity; selecting a corresponding likelihood function according to distribution of fault data ofthe electric energy meter, embedding a B spline function into the Bayesian model, and establishing a hierarchical Bayesian B spline model; after normalization processing is carried out on the effective data with the noise points removed, using the effective data as the input data of the model to train the model; according to the failure data of the electric energy meter, estimating the failure rate of the electric energy meter through the model, so that the problems that an existing data fusion reliability method is incomplete in analysis result, uncertain in model, prone to underfitting undersmall samples and the like, and the accuracy and comprehensiveness are difficult to consider at the same time are solved.