Abstract: To solve the problem of the difficulty in adaptively estimating the fault size of rolling bearings under low-speed and variable-speed compound working conditions, this paper proposes an adaptive estimation method for rolling bearing fault size under low-speed and variable-speed composite working conditions by combining the Savitzky-Golay (SG) filter and Median Absolute Deviation (MAD), based on the Instantaneous Angular Speed (IAS) as the signal source and the characteristics of IAS fluctuations caused by rolling bearing outer rice faults. Firstly, this paper investigates the fluctuation characteristics of IAS caused by rolling bearing outer rice faults under the influence of speed trend components, as well as under different radial loads and different speed working conditions. Secondly, the effectiveness characteristics of the SG filter under different parameter conditions are analyzed, and an adaptive method for determining the optimal parameters of the SG filter is proposed. Finally, based on the adaptive SG filter and combined with the adaptive soft threshold of MAD, the adaptive size estimation of rolling bearings outer rice under low-speed and variable-speed compound working conditions is realized. The effectiveness and robustness of the proposed method were verified by experimental data under different outer rice fault sizes, different speeds, and different radial load conditions.