A two-fluid atomizer has been frequently used in a wide range of industries for various purposes such as painting, cleaning particles and snow making. In particular, the manufacturing of advance semiconductors using sensitive devices such as organic light emitting diodes (OLED) and dynamic random access memory (DRAM), require high performance nozzle. The droplets sprayed with a high relative gas velocity are widely used for cleaning particles. In this paper, two-fluid atomizer is numerically studied according to four variables to confirm the effect on the atomizer performance. The numerical results using the discrete phase model (DPM) with several break-up models are compared with the experimental data measured by the phase doppler particle analyzer (PDPA). Design of experiment (DOE) and genetic algorithm (GA) were used to obtain design points, and conduct sensitivity analysis, respectively. The results showed that the WAVE model has a good agreement compared to the other models, and the orifice diameter is a crucial factor for this model to determine the performance of Weber number and pressure.