The mask spectrometer uses the beam splitter to separate the incident scene into two dimensions: spatial dimension and spectral dimension, then information fusion is used to achieve high spectral resolution of dynamic video signal acquisition. It has wide application value in the field of dynamic hyperspectral imaging. In order to solve the problem of miniaturization of mask spectrometer, for the part of the spectral optical system simplified design of the number of lenses, and the conventional prism was replaced by the grating, the linear dispersion was realized while the structure was further compacted. The problem of stray light generated by the unwanted stages of the grating was analyzed, and the feasibility of the system design was demonstrated. Finally, in the range of 400-1 000 nm, the spectral resolution was less than 4 nm, the average modulation transfer function (MTF) at the full field of view Nyquist frequency was greater than 0.4, and the image illumination uniformity was higher than 0.9, achieved good image quality. Meanwhile, the SNR generated by stray light was 0.06, which did not affect the collection of spectral information.