The stray flux is a commonly used signal in the fault diagnosis of induction motor bearings. However, the weak fault features in the stray flux signal are easily masked by background noise, which greatly increases the difficulty of fault diagnosis. To address this issue, this paper proposes a fault detection method based on maximum envelope spectrum factor deconvolution (MESFD). This method uses the sensitivity of the peak factor of the envelope spectrum to the coupling frequency to suppress the background noise. The initial value of the filter length is determined according to the maximum peak factor of the envelope spectrum, and the filter length is optimized using the snow accumulation and ablation optimizer (SAO). The experimental results verify the effectiveness and superiority of the proposed method in detecting motor bearing faults using stray flux signals.