Heavy-ion fusion reactions between light nuclei such as carbon and oxygen isotopes have been studied becauseof their importance in a wide variety of stellar burning scenarios. However, due to extremely low cross sectionsand signal/background ratio, all the measurements could only be carried out at energies well above the regionof astrophysical interest. The reaction rates in stellar environment could be estimated only by extrapolating theexisted cross sections or the astrophysical S-factors at higher energies. The situation is even more complicated bythe strong, relatively narrow resonances in some reactions, such as 12C+12C, 12C+16O. Traditionally, optical modelor equivalent square-well optical model (ESW) were used to fit the average cross section and predict the reactioncross sections at the energies of astrophysical interest[1]. Recently, a new model, the hindrance model, was proposedto provide systematic fits to fusion reaction data at extreme sub-barrier energies[2]. Lacking of experimental datawithin this energy range, large discrepancies exist among different nuclear reaction models.