Filtering jamming signals from intermediate-frequency data is a widely used method for global navigation satellite system anti-jamming. However, existing filtering algorithms lack signal references to distinguish jamming signals from normal signals, resulting in significant signal distortions caused by filtering out normal signals as jamming. An equalization algorithm based on the frequency-domain adaptive filter (FDAF) is proposed to restore the signal power spectrum after jamming mitigation. First, a signal prediction aided reference spectrum model (SPRSM) is established to provide a local reference spectrum for FDAF. Second, the FDAF equalization model is designed to equalize the filtered signal by using the reference spectrum. Jamming simulation experiments show that the proposed algorithm can improve the carrier-to-noise density ratio and code tracking accuracy while maintaining effective jamming suppression, thereby reducing the signal distortion introduced by conventional notch filtering.