In practical applications, when the coding template of coded aperture spectral imagers does not match the resolution of their detector, the resolution of the system is lowered. For this problem, by using a mathematical model for the Coded Aperture Snapshot Spectral Imaging system (CASSI), its limiting factors such as a mismatch between its coding template and the detector′s resolution are analyzed and the corresponding solutions are given. Considering that the resolution of the coding template is higher than the resolution of the detector, it is proposed that super-resolution technology is introduced to the CASSI system to achieve super-resolution spectral imaging through compressed sensing. For cases in which the resolution of the coding template is lower than the resolution of the detector, a grayscale coding aperture with threshold partitioning grading is proposed to achieve a high-resolution coding mode, which can ensure the resolution of the coded aperture spectral imager. The GPSR algorithm is used to reconstruct the data cube. Experimental results show that the spectral image measured by the CASSI system based on super-resolution theory is more accurate and richer in content. The CASSI system based on a coded aperture with grayscale grading is employed and shown to have higher spatial resolution and spectral resolution. It can be concluded that after optimization, the resolution and imaging quality of the CASSI system are greatly improved and its high resolution components are fully utilized.