For low-temperature combustion (LTC) engines, the ignition and combustion processes are dominantly controlled by the chemical kinetics of fuels. In order to simulate the working process of LTC engines using the multidimensional model, the skeletal or reduced oxidation mechanisms of fuels are urgently required. In this study, a new approach was proposed for the construction and optimization of the skeletal chemical mechanism of primary reference fuel (PRF) by coupling the decoupling methodology with the multiobjective genetic algorithm (GA). An initial skeletal chemical mechanism for PRF was first proposed according to the decoupling methodology. Then, the nondominated sorting genetic algorithm II (NSGA-II) code was employed to optimize the rate constants of the large-molecule reactions in the skeletal PRF mechanism. In NSGA-II, two objective functions based on the experimental data including ignition delay times in shock tubes (ST) and major species profiles in jet-stirred reactors (JSR) were introduced. Compared with the initial mechanism, the performance of the optimized mechanism is improved considerably. Furthermore, the optimized mechanism was used to predict the laminar flame speed, and the major species evolution in premixed laminar flames, as well as the in-cylinder pressure, heat release rate, and emissions in a homogeneous charge compression ignition (HCCI) engine. Good agreements between the predicted and measured results indicate that the integration of the decoupling methodology with genetic algorithm is an efficient approach for the construction of skeletal mechanisms. Furthermore, the influences of the selection and weight factor of the experimental data for mechanism optimization were discussed. The results indicate that the available experimental data on the ignition delay times under extensive operating conditions should be considered in the optimization process. Heavier weight factor should be assigned to the ignition delay times in the negative temperature coefficient (NTC) and low-temperature zones to improve the performance of the optimized mechanism.