Abstract: In order to address the need for coal to underpin energy security under carbon peak ＆ carbon neutrality goals,predictive study of coal prices that affect coal production and supply is presented. Firstly,a long and short-term memory network based on optimized cuckoo search algorithm（CS-LSTM）coal prices forecasting model is set up. The model employs the cuckoo search algorithm to find the optimum for two parameters of the LSTM,namely the learning rate and the number of neurons in the hidden layer,to complete the parameter determination and strengthen the forecasting capability of the LSTM. Secondly,an early warning mechanism for coal prices is established to warn the fluctuations in coal prices. Finally,Shanxi coal price index in 2022 based on CS-LSTM model is predicted,and price warning is provided. The results of the example calculations verify the prediction accuracy of proposed forecasting model and the effectiveness of the early warning mechanism.