The algorithm of spatial heterodyne interferogram matching on finite impulse response and multiple linear regression analysis is proposed based on the radiation transfer model and Beer-Lambert law. For object identification, we extracted the characteristic of potassium double resonance from analyze interferogram directly. Traditional finite impulse response digital filter processing method was adopted to eliminate background information and other interfered information which except from interference frequency information of effective target. By multiple linear regression analysis processing interference data of time-domain of before and after filtering, the optimal coefficients of time-domain filtering had been obtained. So that, interferograms of target signals and background signals were filtered directly. Finally, the signal of potassium double resonance was identified by the interferogram matching. In this algorithm, it’s not required measuring the spectrum and interference pattern of background, neither Fourier transforming of interferogram to access spectrum, and improved efficiency and increased inverting capacity of spatial heterodyne spectroscopy analysis remote sensing data simultaneously.