Human activities, primarily fossil fuel burning, have increased the amount of carbon dioxide in the atmosphere and absorbed by the ocean. The excess carbon dioxide absorbed by the ocean leads to ocean acidification. Coastal acidification along the North American East Coast is not well understood due to limited spatial and temporal observations and because other factors besides atmospheric uptake affect coastal acidification. In this study, we use a machine learning approach to recognize the surface water carbon chemistry patterns in the Mid-Atlantic Bight and the South Atlantic Bight, which are two coastal regions of eastern North America, and then use the trained algorithms to fill in the large observational gaps. The reconstructed time series data show that the pH of coastal water in the Mid-Atlantic Bight was lower and the rate of acidification was faster than that of the South Atlantic Bight. The rate of surface water acidification was faster in winter when pH was the lowest, thereby exerting more stress on shell-building organisms in the Mid-Atlantic Bight.