The numerous factors influencing the air-sea carbon dioxide(CO_2) transfer velocity have been discussed for many years, yet the contributions of various factors have undergone little quantitative estimation. To better understand the mechanism of air-sea transfer, the effects of different factors are discussed on the air-sea transfer velocity and the various parametric models describing the phenomenon are classified and compared.Then, based on GAS EX-98 and ASGAMAGE data, wind models are evaluated and the effects of some factors are discussed quantitatively, including bubbles, waves, wind and so on by considering their interaction through a piecewise average approach. It is found that the air-sea CO_2 transfer velocity is not only the function of the wind speed, but is also affected by bubbles, wave parameters and other factors. Stepwise and linear regressions are used. When considering the wind speed, bubbles mediated and the significant wave height, the root mean square error is reduced from 34.53 cm/h to 16.96 cm/h. Discussing the various factors quantitatively can be useful in future assessments of a large spatial scale and long-term air-sea CO_2 flux and global change. The numerous factors influencing the air-sea carbon dioxide (CO_2) transfer velocity have been discussed for many years, yet the contributions of various factors have undergone little quantitative estimation. To better understand the mechanism of air-sea transfer, the effects of different factors are discussed on the air-sea transfer velocity and the various parametric models describing the phenomenon are classified and compared. Then, based on GAS EX-98 and ASGAMAGE data, wind models are evaluated and the effects of some factors are discussed quantitatively, including , waves, wind and so on by considering their interaction through a piecewise average approach. It is found that the air-sea CO_2 transfer velocity is not only the function of the wind speed, but also also affected by bubbles, wave parameters and other factors . Stepwise and linear regressions are used. When considering the wind speed, bubbles mediated and the significant wave height, the root mean square error is reduc ed from 34.53 cm / h to 16.96 cm / h. Discussing the various factors quantitatively can be useful in future assessments of a large spatial scale and long-term air-sea CO_2 flux and global change.