In this paper, a simulation-based optimization methodology for nanoscale complementary metal-oxide-semiconductor (CMOS) device fabrication is advanced. Fluctuation of electrical characteristics is simultaneously considered and minimized in the optimization procedure. Integration of device and process simulation is implemented to evaluate device performances, where the hybrid intelligent approach enables us to extract optimal recipes which are subject to targeted device specification. Production of CMOS devices now enters the technology node of 65 nm; therefore, random-dopant-induced characteristic fluctuation should be minimized when a set of fabrication parameters is suggested. Verification of the optimization methodology is tested and performed for the 65-nm CMOS device. Compared with realistic fabricated and measured data, this approach can achieve the device characteristics; e.g., for the explored 65-nm n-type MOS field effect transistor, the on-state current > 0.35 mA/mum, the off-state current < 1.5e - 11 A/mum, and the threshold voltage = 0.43 V. Meanwhile, it reduces the threshold voltage fluctuation (sigmavth ~ 0.017 V). This approach provides an alternative to accelerate the tuning of process parameters and benefits manufacturing of nanoscale CMOS devices.