The results show that the average level of carbon price in each pilot market is quite different. The data of each carbon market show the characteristics of peak thick tail distribution, volatility concentration and multifractal. In the process of analyzing the monthly average price of carbon in China, it is found that the establishment of a new carbon emission market will improve the transaction price of each carbon market, and the proper adjustment of the free carbon emission quota proportion will help to reduce the transaction price of carbon emission quota, and the gradual improvement of the accounting and verification system of carbon emission market will make the transaction price of carbon emission quota stable. At present, the GARCH family model commonly used can not reflect multifractal characteristics, which is easy to lead to estimation bias. In this paper, Markov transform multifractal (MSM) model is used to predict and analyze the carbon market, and the result with high prediction accuracy is obtained, which is of reference significance for policy-making and the formation of national carbon emission price.