In the structure of China’s energy system, coal has always occupied a dominant position. However, at present, the carbon emission caused by coal combustion accounts for about 60% of the total carbon emission in the country, and the unavoidable consequence of its large-scale use is the environmental pollution, so it is urgent to adjust the energy structure. Firstly, we analyse the forecast of coal demand and the market demand for coal industry capacity removal in China every five years until 2050. Due to the large amount of forecast data, we establish ARIMA model for forecasting. For the prediction of coal demand, we firstly carry out the smoothness test on the data we collected, judge whether it is smooth by the P value, and then set up the optimal ARIMA model to get the prediction result, that is, the demand for coal in the next 30 years is still showing the pattern of steady increase, but the growth rate has been slowing down. For the prediction of market demand for coal industry capacity removal, we use the production coefficient elasticity prediction method to establish a prediction model. After that, we use the moving average method and ARIMA model to predict the elasticity coefficient and GDP growth rate in turn, and then we get the results to calculate the coal demand and the market demand of the industry’s decapacity according to the prediction model, and we can know that our country can completely realise the decapacity of the coal industry roughly in the period of 2026–2030.