According to evolution law of carbon emission, by using the combined model from ARIMA Model and BP neural network integration, this paper studies the prediction of carbon emission quantity of China. By taking carbon emission quantity of China during 19802007 as training sample, the model parameters are made, then the combined model in this paper is tested by taking the carbon emission quantity of China during 20082013 as testing sample and the model is compared with the prediction model established in the existed references. The result shows that the deviation is relatively small by using the combined model constructed in this paper, and finally 2014—2020 carbon emission quantity of China is predicted based on the combined model and the predicted result show that carbon emission quantity of China will continue to face reduction pressure.