The total radiance observed by satellite-borne ocean color sensors results from several contributions: the atmospheric molecular and aerosol scattering, the sea surface reflectance and the water leaving radiance. In order to apply these ocean color sensor data to access the ocean parameters, atmospheric correction should be undertaken in advance to extract the water leaving radiance that is relevant to the ocean parameters. The aim of this study is to establish one atmospheric correction process for the Ocean Color Imager (OCI) of ROCSAT-1, and some pre-launch simulation results are demonstrated. The simulation shows that the water leaving radiance estimation seems to be reliable by the ocean color radiance algorithm proposed in this paper, because the estimated total pigment concentration seems to be reasonable. Further investigations done in this study using SeaWiFS data reveal that the water leaving radiance is somewhat overestimated in comparison with ship measurements because of the underestimated aerosol scattering. The reason for this underestimation is probably the inconsistent marine aerosol type and size distribution between the real and input atmospheric parameters. The simulated result also shows the sun glint effect is significant and plays the most important role in the radiance contributions of OCI imagery. Consequently, choosing the most suitable acquiring time of OCI in order to avoid the sun glint effect will be very important in future OCI operation scheduling.