In this study, an adaptive multiple-gain imaging method was proposed based on a pushbroom imaging spectrometer to meet the demands of high dynamic range (DR) monitoring using an ocean color remote sensor. Further, the advantage of the pixel-by-pixel adaptive readout method was explained by analyzing the differences in case of DR and the signal-to-noise ratio (SNR), and the corresponding system design was presented. The limit of full well charge was determined based on the ocean imaging scene, and the optimal number and proportion of gains were obtained via system simulation based on the constraints associated with respect to SNR and DR. The results denote that the optimal imaging effect can be obtained when the number of gains is four and the full well charges are 2.5 Me ^- (ultra-low gain), 500 ke ^- (low gain), 100 ke ^- (medium gain), and 20 ke ^- (high gain). Based on this setting, the total DR of the spectral imaging system can become 116 dB; the average SNR of each channel under typical radiance becomes 767.37. Furthermore, the medium-resolution atmospheric radiation transfer model was used to verify this imaging mode. The conditions with respect to the channels meet the requirements of DR and SNR, except for the 10th channel of the cloud that appears to be out of saturation.