Prestressed concrete bridges (PCBs) are integral components of infrastructure, and accurate identification of internal forces (shear force and bending moment) within the bridge is vital to their design, performance and maintenance. The streamlined method for identifying internal forces at bridge interfaces is particularly important when considering the reduction of prestress forces (PFs) over the lifetime of the PCB and the integration of various moving forces (MFs). This approach is essential to guarantee the bridge’s structural soundness and adequacy. A substructural modelling technique is introduced, which involves dividing the PCB into multiple segments (substructure). This kind of modelling aims to decrease the complexity associated with identifying internal forces without analysing the entire bridge. A Lagrangian polynomial-based Hermitian interpolation method is incorporated to convert the multiple MFs into equivalent vertical forces at the interface of each substructure (nodal forces). A finite element-based approach has been proposed to identify internal forces by integrating nodal forces caused by multiple MFs and equivalent external pseudo forces caused by PFs. In addition, the method incorporates inertia forces and damping forces to overcome the challenges associated with conventional methods of internal force identification. Based on experimental data from various locations along a simply supported prestressed box-girder bridge, the proposed method for internal force identification has been verified, with a maximum relative percentage error of 6.77%. The analysis demonstrates that this method can accurately and efficiently identify internal forces using a reduced number of computational resources cost-effectively. Ultimately, this study presents a simplified methodology for identifying internal forces in PCBs when multiple MFs are present. Furthermore, by obtaining the internal forces, bridge managers can use the proposed method to identify PF to ensure the structural soundness of PCBs with a limited number of sensors and less computational effort.