Abstract: To address the limitations of conventional body-fitted grid methods—such as difficulties in mesh generation, low computational efficiency, and poor robustness in complex fluid-structure interaction (FSI) simulations—this paper develops an efficient numerical framework based on the lattice Boltzmann-immersed boundary method (LB-IBM). In this framework, the lattice Boltzmann method (LBM) is employed to solve the incompressible Navier-Stokes equations, the finite difference method (FDM) is used to discretize the structural dynamics of flexible bodies, and a fourth-order hybrid linear multi-step scheme is adopted to handle rigid-body vibrations. Fluid-structure coupling is achieved through a penalty-based immersed boundary method (IBM), with the amplitude prediction error controlled within 5% in rigid-body vortex-induced vibration cases. Additionally, the framework integrates multi-block grid refinement and hybrid MPI/OpenMP parallel computing to support large-scale high-efficiency simulations. Typical benchmark tests show that the mesh count is reduced by approximately 40%, while the errors in key parameters such as the Strouhal number and drag coefficient remain below 5%. The framework is validated and applied through representative cases, including bluff body flows, flow-induced vibrations, and bio-inspired dragonfly flapping. Results demonstrate that the proposed approach effectively overcomes the limitations of conventional body-fitted methods, exhibiting good computational efficiency, reliability, and robustness in simulating FSI problems characterized by complex geometries, large displacements, large deformations, or multi-body motions.