To ensure the software reliability of smart meter is beneficial to reduce the failure rate of equipment and improve the service quality of electric power marketing department. The software testing technology based on grey box dynamic testing can check the software source code safely and deeply. Based on the software testing platform, this paper constructs several Markov states according to the probability of event trigger, and carries on Monte Carlo simulation analysis to the smart meter. Then, the software reliability of two single-phase smart meters is investigated based on the proposed test method. The simulation results verify that the proposed test model can fully ex-cavate the software maturity and robustness of the smart meter, and effectively generate contrast data and conclusions.