This paper introduces an optimized orbit jump strategy for nonlinear vibration energy harvesters (VEHs). Nonlinear VEHs are a promising alternative to linear VEHs due to their broadband characteristics. However, they exhibit complex dynamical behaviors, including not only high-power inter-well orbits but also low-power intra-well orbits and chaos. The existence of low-power orbits in their dynamics can restrict their energy harvesting performance. In order to overcome this issue, this study investigates an orbit jump strategy that allows the VEH to transition from low-power intrawell orbits to high-power inter-well orbits. The orbit jump strategy, which is based on varying the buckling level of a bistable VEH, has been previously studied but not yet optimized. In this study, we optimize this orbit jump strategy to ensure practical reproducibility and robustness against variations in parameters or excitation. Through the combination of thorough experimental identification and high-performance computing of the complex transients during orbit jumps, we achieved high numerical accuracy in orbit jump modeling. This was possible by a developed Python CUDA code using GPU parallel computing to handle a large number of