The spatial simulation scale of forest carbon cycle process models is diverse but the data acquisition is difficult. It is impossible to install a carbon flux observation system in different environments in the study of forest carbon change response and adaptation. In this paper the remote sensing process coupling model was proposed to quantitatively estimate the dynamic GPP value NPP value RH value and NEP value to invert the carbon sink / carbon source situation of different dominant tree species of forest vegetation community to explore the ecological value of vegetation restoration measures such as afforestation forest transformation and forest artificial management and to activate the reconstruction effect of carbon source effect. Through the combination of field survey sampling method with space instead of time and remote sensing model the driving factors of temperature influence coefficient evapotranspiration water-pressure influence coefficient carbon dioxide concentration stress plant respiration and soil respiration in GLOPEM-CEVSA model were input to study the differences of forest carbon storages at the forest patch scale. The research showed that 1 in the picture as a whole Net Primary Productivity of Woodland>Grassland>Agricultural Land>Water Area>Construction Land Net Ecosystem Productivity of Woodland >Agricultural Land >Grassland >Construction land >Water Area 2 from the inter-annual NPP / NEP values most forest ecosystems showed atmospheric CO2 carbon sink 3 the relationship of NPP values in the natural restoration succession of forest community was Deciduous Broad-Leaved Forest Community A4 > Pinus Massoniana / Cunninghamia Lanceolata Evergreen Coniferous Forest Community A5 >Cypress Evergreen Coniferous Forest Community A6 >Evergreen Shrub / Shrub Grass/ Artificial Shrub Community A2 >Herbaceous Community A1 NEP values were as follows Deciduous Broad-Leaved Forest Community > Pinus Massoniana / Cunninghamia Lanceolata Evergreen Coniferous Forest Community >Evergreen Shrub / Shrub Grass/ Artificial Shrub Community >Cypress Evergreen Coniferous Forest Community>Herbaceous Community the relationship of NPP values in artificial assisted succession was as follows Bamboo Forest Community > Pinus Massoniana / Cunninghamia Lanceolata Evergreen Coniferous Forest Community>Cypress Evergreen Coniferous Forest Community>Evergreen Shrub / Shrub Grass/ Artificial Shrub Community NEP values were as follows Bamboo Forest Community>Pinus Massoniana / Cunninghamia Lanceolata Evergreen Coniferous Forest Community > Evergreen Shrub / Shrub Grass/ Artificial Shrub Community > Cypress Evergreen Coniferous Forest Community and both the NPP values and the NEP values showed a unimodal development trend of rising first and then decreasing with the sequential succession.