Real-time prediction of wireless link quality is a main problem which needs to be solved in the smart grid station area to guarantee the transmission reliability of wireless communication applications. The superposition of nonlinear and non-stationary sequences expressed by signal to noise ratio(SNR)time series of wireless link quality is the main factor that affects the prediction accuracy. For this reason

reliability confidence interval prediction is proposed by approximately decoupling the wireless link SNR sequence into nonlinear and non-stationary random sequences. And the wavelet neural network is used to develop the wireless link SNR sequence prediction model and the variance of non-stationary random sequence prediction model. Based on the model prediction results

this paper calculates the upper bound and lower bound of the wireless link reliability confidence interval. In the actual smart grid environment

the correctness and effectiveness of the proposed algorithm are verified

and the practicality of the proposed algorithm is illustrated through an application example. This work is supported by National Natural Science Foundation of China(No. 51307041).