The cross-correlation between carbon spot and futures markets reflects the risk conduction mechanism between the two markets. Deeply depicting and analyzing this risk conduction mechanism is of great significance for investors to carry out risk management strategies. Considering the nonlinear and asymmetric characteristics of cross-correlation between carbon spot and futures markets, this paper applies multifractal cross-correlation analysis method to investigate the cross-correlation between carbon spot and futures markets. Firstly, through Empirical Mode Decomposition (EMD)-Multifractal Detrended Cross-Correlation Analysis (MF-DCCA) detection, it is found that there is an obvious cross-correlation between carbon spot and futures markets, and the cross-correlation has multifractal characteristic. Secondly, by using EMD-time delay-DCCA method, we find the conduction direction between carbon spot and futures markets is bidirectional, and the futures market has a greater impact on the spot market in the short term. Thirdly, through using EMD-MF-ADCCA method, we find the cross-correlation between the two markets is asymmetric, and the cross-correlation between the two markets is more significant when carbon market is in a downward trend than in an upward trend. Fourthly, through constructing EMD-time delay-ADCCA model, we find that there is a two-way asymmetric conduction effect between carbon spot and futures markets when the lag period is short, and when the carbon market is in a downward trend, the conduction effect between the two markets is stronger. However, with the extension of time lag, the conduction effect of the two markets no longer presents obvious asymmetric characteristics.