Abstract: This paper proposed a short-term solar irradiation forecast model based on long/short term memory neural network. The training samples with recursive structure are used to guarantee the time correlation within training sample. A prediction model based on long-short-term memory neural network is established and compared with the traditional prediction model based on artificial neural network. The results shown that compared with the traditional prediction model based on artificial neural network,the prediction model based on long-short-term memory neural network can significantly reduce the mean square error,which indicates that the model is more suitable for GHI prediction.