In order to deal with the smart electric meters’ mass information, varied fault types and sudden faults, the fault and life prediction model for smart electric meter based on random forest is established. Based on the theory of large data analysis, this paper excavates the accumulative data of large numbers of smart electric meters and forecast its fault and life, compare with other models. The experimental results demonstrate the effectiveness and correctness of the proposed model. It can provide reference for data mining in the research of smart electric meter.