For high-dimensional complex combustion reaction kinetic systems, conventional simplification methods based on elementary reaction approaches (e.g., sensitivity analysis [SA]) struggle to derive compact reduced mechanisms due to the nonlinear characteristics of the kinetic systems, leading to the widespread adoption of graph-based methods in mechanism reduction. To overcome the dual challenges of reducing high-dimensional stiff mechanisms and resolving the inherent limitations of elementary reaction-based simplification methodologies, a novel mechanism reduction framework employing an improved binary genetic algorithm (IBGA) was established. The IBGA operates through binary-encoded particles where each bit corresponds to an elementary reaction: 0 indicates exclusion from the simplified mechanism, while 1 denotes retention. The optimization objective maximizes the number of zero-value bits while preserving critical combustion characteristics. This methodology was implemented for the mechanism reduction of ethylene and dimethyl ether (DME) combustion systems, etc. Results indicate that the IBGA-based approach achieves significant mechanism size reduction while maintaining accuracy. The ethylene mechanism was reduced to 28 reactions, and the DME mechanism to 40 reactions. Furthermore, in order to further validate the performance of the IBGA reduction method, the ethylene mechanism and DME mechanism are also reduced for predicting both ignition delay time and laminar flame speed. The results shown C2H4/air reduced mechanism involving 28 species and 56 reactions and DME/air reduced mechanism involving 31 species and 92 reactions are obtained. The obtained simplified mechanisms exhibit enhanced compactness with preserved prediction fidelity for combustion characteristics. A comparative analysis between the IBGA and deep mechanism reduction (DeePMR) methods in reducing high-temperature LLNL butanol isomers’ mechanisms demonstrates that the IBGA significantly shortens the required computational time for reduction while producing more compact mechanisms.