Due to the reliability evaluation result based on a single data source is not accurate enough caused by the field data of the smart meters not covering its entire life cycle and the difference between the stress of accelerated life test and the actual operating environment, a reliability assessment method of smart meters combining Bayes and Bootstrap methods is proposed. The proposed method uses Bootstrap method to process the field data to obtain the discrete distribution of the reliability model parameters of the smart meters, then adopts the Bayes method to fuse the discrete distribution, as prior information, and the accelerated life test data to obtain the parameter estimates, and realizes the reliability assessment of smart meters. The result of the example shows that the reliability model of the smart meters obtained by the proposed method approximates the actual field situation in the first half and is supported by test data in the second half, which can be used for the reliability assessment of the entire life cycle of the smart meters.