Abstract: In the neutralizer of the magnetic confinement fusion neutral beam heating system, the gas target is the medium for the ion beam to be converted to a neutral state, and sufficient gas target thickness can make the ion beam achieve the optimal neutralization efficiency. However, this gas is used as gas load in the vacuum chamber, and it is hoped that the less the better. In order to optimize the flow required for the optimal target thickness, it is expected to obtain the optimal target thickness with less gas flow. According to the molecular flow characteristics of the gas in the neutralizer, a code based on the test particle Monte Carlo method was used to simulate the distribution of the molecules in the neutralizer. In this code, the emission angle and flight velocity of molecules in the collision reaction were taken as the objects to study the molecular density, and the integral of molecular density in the length direction of the neutralizer reflected the size of the target thickness, so as to obtain the optimization scheme. The results show that the emission angle of the molecule after collision is random, but the neutralizer gas supply angle can affect the target thickness. The angle between the direction of the gas supply and the direction of the beam is the best at 150°, the worst at 15°, and it is consistent with the law of the average number of collisions of molecules in the neutralizer. After each collision, the molecule gets the temperature of the wall, which makes its flight velocity change, so the neutralizer wall temperature will also affect the target thickness. The same flow can obtain a larger target thickness at a lower temperature. In particular, when the temperature is below 100 K, the influence of temperature on the target thickness is more obvious, and the slope of linear fitting is 5 times that of higher than 100 K. The temperature also causes a change in the time to establish the optimal target thickness, and the lower the temperature, the longer the time. The results provide theoretical basis and data support for the optimal design of neutralizer structure and experimental control.