The flutter control effect and mechanism of central-slotting, which have gradually been adopted in the design and construction of long-span bridges as an effective flutter controlling measure, were investigated with theoretical analysis and wind tunnel test. Five basic girder cross-sections representing five typical aerodynamic configurations were selected and central-slotted with two different slot widths. Then, a series of sectional model tests and theoretical analyses based on the two-dimensional three-degrees-of-freedom coupling flutter analysis method (2 dimension-3 degrees of freedom method, 2d-3DOF method) were carried out to investigate the aerodynamic performance, flutter mechanism and flutter modality of the five basic sections and their corresponding central-slotted sections. The results show that central-slotting can not always improve the aerodynamic stability of bridge structure. The control effect of central-slotting depends on the aerodynamic configuration of the original girder section and the corresponding central-slotting width. If the original section is inappropriate or the slot width is unsuitable, central-slotting will even deteriorate the structural flutter performance. Theoretical investigations indicated that the differences in flutter control effects come from the different formation and evolution of aerodynamic damping, and flutter modality especially the participation level of heaving motion also has a significant influence on the control effect of central-slotting.