<p>A study is conducted with simulation models to optimize the stability of gaseous pressure in a compressed natural gas (CNG) injection system. The simulation models are established for the injection system, including an operating model of a rail pipe, mechanical and electromagnetic models of a CNG injector and a pressure regulator. An ON–OFF controller is applied for the pressure regulator to keep a stable gaseous pressure in rail pipe as the injection processes is activated. The structural parameters of CNG injection system are selected to investigate their effects on the gaseous pressure stability in rail pipe, including volume of rail pipe, spring stiffness and number of coil turns in the pressure regulator. To optimize the stability of gaseous pressure in rail pipe based on above structural parameters, a particle swarm optimization (PSO) algorithm is applied. The study results indicate that the gaseous pressure stability in rail pipe could be improved when increasing rail pipe volume, increasing spring stiffness and reducing number of coil turns in the pressure regulator. PSO algorithm allows to find quickly the best structural parameters to optimize the gaseous pressure stability in rail pipe during injection processes.</p>