Explaining the formation and synergistic effect in multi-active sites is a vital factor in designing a highly efficient catalyst. In this study, a series of new graphene oxide(GO) supported arylimine Pd(II), Pd(II)/Er(III), and Pd(II)/Cu(II), Pd(II)/Er(III)/Cu(II) (GO@A-Pchmba-Pd, GO@A-Pchmba-Pd0.1Er0.9, GO@A-Pchmba-Pd0.1Cu0.9, GO@A-Pchmba-Pd0.1Er0.4Cu0.5) catalytic films were fabricated using self-assembly. Among these catalysts, GO@A-Pchmba-Pd0.1Er0.4Cu0.5 exhibited the highest activity (TOF = 145,506 h−1) in Suzuki cross-coupling reaction with a good compatibility, but poor recycling stability. The catalytic film acted not only as a "reservoir" to supply the active sites, but "stabilizer" to stabilize the active species, which provided a research platform to gain deep insight into the truly active center. The synergistic effect in the truly active center containing Pd/PdO/Cu2O/Er2O3/PdCu3 was described as that: the electron transferred from GO to Pd/Er/Cu via ligand, and then the electron transferred from erbium to palladium or its oxide, or from erbium to copper or cuprous oxide and then to palladium, resulting in enhancing the more electronegative Pd and further promoting the activity. The aggregation of the active metals, the loss of the active metals, residues absorbed on the surface, some destroyed organometallic in film, and the reduced average pore size were the crucial reasons for the deactivation and instability.