Aiming at the problem that it is difficult to describe the degradation trend of smart meters under typical environmental stresses, such as high dry heat, high altitude and high cold, a method of smart meter degradation based on multi-fusion Bayesian zero-value wiener was proposed. Firstly, the multi-dimensional analysis of the basic error and the distribution law of environmental stresses in the samples of several smart meters of multiple company is carried out to select a reasonable prior distribution. Then, the multi-fusion Bayesian network model is used to integrate the influence of various environmental stresses on the degradation trend of various smart meters of each company, and then a variable zero-value wiener degradation model is proposed. After solving the model parameters under non-conjugated conditions by using markov chain monte carlo method, the applicability and superiority of the variable zero-value wiener degradation model is verified by multiple models’ fittings, predictions comparison analysis and its confidence interval. Finally, the reliability of the smart meters is analyzed with the inverse gaussian distribution. The results show that the model can accurately analyze the degradation trend of the smart meters with time under typical environmental stresses, and it also shows that it is suitable for the prediction of the basic error of the smart meters under typical environmental conditions.