By using doppler asymmetric spatial heterodyne spectroscopy and doppler effect, the wind speed can be achieved through detecting the interferogram of airglow in the upper atmosphere. This paper mainly analyses the data processing method of the interferogram and then derive the interferometer phase in order to get the wind speed. Comparing with the traditional spatial heterodyne spectroscopy, not only the noise and error of the system should be taken into consideration, but the window function that used to isolate the spectrum has a great influence during the data processing. Then the effect of window type and window width on phase difference of interferogram and the wind error curve are simulated through software. On basis of this the wind error curve under the noise of system and flat field factor are simulated by choosing appropriate window function. The window function simulation indicates that although the joining of window leads to a distortion of the interferogam and phase, the wind speed error can be less than 0.5% with Hanning window in the appropriate optical path difference. The noise of the system simulation indicates that the wind speed error increases with the noise, so it is necessary to control the system noise and preprocess the sampling data. The research on data processing method has great theoretical significance and practical value for designing the system parameter and improving the precision of spatial heterodyne wind detection.