To address the issues of insufficient spatial resolution and data missing caused by cloud occlusion in geostationary satellite ocean color data, this study proposes a super-resolution reconstruction method that integrates dynamic cloud masking training. Focusing on the chlorophyll a concentration product of the Himawari-8 satellite, this method aims to utilize low-resolution (0.05°) data covering the entire Western Pacific to simultaneously generate a complete dataset with high resolution (0.01°) and no cloud missing. A dynamic composite masking strategy is embedded into the training process of the ESPCN (Efficient sub-pixel convolutional neural network). It enables the model to learn the ability to recover complete spatial structures from partially obscured observations while achieving 5× super-resolution reconstruction. Experimental results show that the model achieved good reconstruction accuracy on the test set (Peak signal-to-noise ratio>50 dB, r>0.86). Based on this method, this study produced a high-resolution, spatiotemporally continuous chlorophyll a dataset covering the entire Western Pacific from 2015 to 2022. Rigorous validation with independent Argo float data shows that the consistency between the reconstructed data and in-situ observations is significantly better than that of the original low-resolution product. The number of matching samples increases from 3 783 to 5 141, the root mean square error (RMSE) decreases from 0.693 to 0.289, and the mean absolute error (MAE) decreases from 0.203 to 0.171. At the application level, the reconstructed dataset can effectively capture fine-scale oceanic features in typical coastal areas, such as the Pearl River Estuary. It can also clearly delineate the estuarine plume front structure and the "low-high-low" parabolic evolution pattern of chlorophyll a concentration. This study provides a new technical pathway for producing high-quality marine remote sensing data products and offers critical data support for researching on the evolution of coastal water quality.