Accurate quantification of infrared remote sensing signal is important for acquisition of pollutant cloud’s information, but spectral distortions occurred in measurement may hinder the achievement of such purpose. An adaptive method based on instrumental line shape (ILS) model was established in order to compensate the contributions due to ILS distortion. Through analysis of the sources of ILS function, the ideal, inherent function as well as phase error contribution were modeled based on design parameters of a real infrared spectrometer. Furthermore, an algorithm which reconstructs ILS function from measurement was developed by using iterative optimization method, which takes root mean square between differences of simulation and measurement spectrum as cost function. The compensation result by using reconstructed ILS function on simulated spectrum suggests that differences between simulation and measurement were effectively eliminated. The analysis showed that inherent ILS may cause spectral feature broadening toward low frequency, and phase error is responsible for spectral feature asymmetry. All three sources of ILS distortion must be considered simultaneously to get accurate pollutant cloud parameter from measured spectrum. The acquisition of distortion parameters and the corresponding compensation method may be helpful for the recognition and quantification of infrared remote sensing signals.