Abstract: The spatial heterodyne spectrometer is a new ultra-high resolution spectrometer, which can be used for the atmospheric monitoring, satellite remote sensing and other fields. In order to reduce the noise in the spatial heterodyne spectral signal, a denoising method based on lifting wavelet transform combined with median filter was proposed. The two-factor threshold function and hierarchical threshold selection were incorporated by the improved lifting wavelet transform. Compared with the soft and hard thresholds of the wavelet transform, the spatial heterodyne spectrum could be extracted, the peak width could be reduced, and the important detail features could be preserved by the proposed algorithm, which had superior denoising effect than the wavelet transform. Finally, the two quantitative indexes of SNR and mean square error were used to measure the effect of the proposed algorithm. The experimental results show that, compared with the soft threshold method, the signal-to-noise ratio of the algorithm for the xenon lamp and the integrating sphere is increased by 24.6% and 31%, respectively, and the mean square error is reduced by 43.2% and 51.5%, respectively. Compared with the hard threshold method, the signal-to-noise ratio is increased by 21.5% and 30.6%, respectively, and the mean square error is reduced by 40.2% and 51.2% respectively. Therefore, the algorithm is feasible in the spatial heterodyne spectral noise reduction.