CO2 is one the crucial green house gases. Oceans act as the major sink of atmospheric CO2. Derive the sea surface pCO2 with site measurements in the field is difficult to implement, labour-costly and limited to a few sites. Images acquired by space-borne satellite are of great capability to capture the real time water properties including chl-a over a wide spread of hundred kilometres with continuous measurements at a few hundred meters. It is also vital to investigate the spatial and temporal evolution of sea surface pCO2 over a certain area of interest. In this study, we aimed to derive the sea surface pCO2 over the Baltic Sea from 2002 to the present on monthly scale with ~10 sea surface properties derived from satellite images. As the data set are relatively big, developing the effective algorithms involving machine learning would take much more computing power than what a single desktop can provide.