The smart electricity meter is an important and sophisticated metrical equipment in smart grid. Its fault types have complicated characteristics of randomness and fuzziness, which have significant impacts on electrical safety of consumer as well as fairness and justness of measurement directly. Therefore, a classification diagnosis cloud model of fault information of smart electricity meter is proposed in order to analyze the randomness and fuzziness of fault data in smart electricity meter and their correlation. In this model, the clustering algorithm is applied, and the cloud theory is employed to conduct data mining. In addition, the weakness and key parts affecting reliable operation of smart electricity meter is also specified. Finally, the abnormal event record of smart electricity meter in a certain area in 2015 is chosen as the data source for analysis. The effectiveness and practicability of the proposed method is verified.