For the problem that the weak features of rolling bearing compound failure mode are difficult to extract， a method for rolling bearing compound fault diagnosis based on adaptive iterative filtering （AIF） and improved time-time （TT） transform is proposed. First， AIF method is applied to decompose the fault vibration signal to achieve a series of intrinsic mode functions and the maximum correlated kurtosis criterion is adopted to select the characteristic components. Then， the improved TT transform method is used to denoise the characteristic components to reinforce the impact features. Finally， the denoised characteristic components are performed on envelope analysis to extract the fault characteristic frequencies and complete the identification of rolling bearing fault mode. The simulated test and actual fault diagnosis instance show that the method proposed in this article can effectively identify the weak fault feature message of the fault signals.