Radio-Frequency Interference (RFI) adversely affects useful signals, thereby seriously affecting image quality. In this study, a modified eigensubspace-based approach for radio-frequency interference suppression of Synthetic Aperture Radar (SAR) images is proposed. In the preprocessing stage of our proposed algorithm, RFI detection is conducted in both the frequency and time domains. Subsequently, we can only deal with the data containing the RFI via the traditional eigensubspace-based approach. Compared with the traditional eigensubspace-based approach, our proposed algorithm can function more efficiently and effectively.