In light of the deficiency in the existing researches, this paper uses the OSW-MF-DFA and OSW-A-MFDFA to reconstruct a risk model from the perspective of multifractality. The long memory, risk and efficiency of the bull and bear of CSI300 index futures from 2014 to 2016 are used as the objectives of study. In this paper, we analyze the problem from three aspects:traditional paradigm, asymmetry and stage, based on the high-frequency data. The empirical results show that the return series of bull market, bear market and stages are multifractal, meanwhile, they are asymmetric; there are significant divergence among the bull market, bear market and stages in terms of their long memory, risk and efficiency, and the asymmetry within them differs from each other, which is largely due to different economic background, market policy and investor behavior; in addition, the bull and bear markets have their own stage characteristics respectively. This research has practical significance for regulatory policy formulation, risk management, investment strategy building, and so on.