The lubrication system is one of the most important subsystems in gasoline internal combustion engines (ICEs), which provides hydrodynamic lubrication for friction pairs. The performance of the lubrication system affects the performance of the engine directly. The objective of this work is to reduce the friction loss of the engine and the driven power of the oil pump through design optimization. Two most important oil consumers in the lubrication system are investigated using multibody dynamics (MBD) and elastohydrodynamics (EHD). Considering that MBD and EHD analyses are time-consuming, Kriging is applied to establish the approximation models for bearings. Multi-objective optimization of bearings based on approximation models is formulated and conducted. Given the difference among multiple cylinders in the engine, a bilevel optimization framework is used to perform bearing optimization. The oil consumption and the friction loss of the bearings are reduced within the entire speed range. After that, the pipe diameters of the lubrication system are optimized with optimized bearings to reduce the flow resistance. With the optimization of both bearings and lubrication pipes in a sequential manner, the oil pressure is maintained at the baseline level while the oil pump size is reduced, and the driven power is averagely dropped over the entire speed range.