To date, the measurements from AIRS, GOSAT, SCIAMACHY and IASI are widely used to derive atmospheric CO2 concentration globally. Quantification of the spatio-temporal differences between these CO2 products is crucial for deeply understanding each product and for exploring the potential for their joint use in the future. A series of processing has been proposed in this paper to unify these existing CO2 products, so that they can be theoretically comparable. Based on this, the spatio-temporal variations of these CO2 products have been conducted. The results reveal that AIRS show the largest spatial coverage of CO2 compared to other CO2 products. The valid XCO2 data from SCIAMACHY are mainly restricted to land regions. Even if over the land, the XCO2 distribution of ACOS is slightly wider than that of SCIAMACHY, especially over the region of Eurasia and North America. An obvious arch-shaped pattern along latitude for ACOS, Japan-GOSAT and SCIAMACHY is detected, while no distinct latitudinal variation can be observed for AIRS. For the seasonal variation, these datasets show a similar trend with the maximum CO2 loading occurring in spring. The discrepancy between these products implies that it is greatly necessary to better constrain the uncertainties in CO2 retrieval from space in the future.