The formation mechanism and evolution trend of different failure modes on-site operation electric energy meter are different. This paper proposes a method for predicting the life of electric energy meter based on multiple failure Weibull distribution model. Using historical fault data of electric energy meter to establish Weibull distribution model of failure rate of each failure mode. According to the goodness of fit, we can optimize the prediction results of some failure modes and accumulate the stage failure rates of all failure modes, then the expected life of the overall electric energy meter is obtained. By verifying the actual energy meter fault data on the site, showing that the method can effectively predict the life of the electric energy meter, discover potential faults in advance, and provide technical support for batch electric energy meter status replacement and hidden danger investigation.