Groove parameters in gas film seal with grooved interface make an obvious impact on the performance of seal system. There are many parameters to describe the geometric features of the groove. In general, a big limitation exists in one-dimensional optimization of groove geometry. Based on particle swarm intelligence algorithm, this article proposed and carried out multi-dimensional optimization of groove geometry in gas film seal, regarded the groove geometry parameters as components of the particle, completed numerical solution of the objective function for the seal performance, and obtained better groove geometry parameters. The example showed that the effect that every geometry dimension plays on the steady-state characteristics of cylindrical gas film seal is not independent, and the multi-dimensional optimization method effectively improves the results of the objective function value. The proposed method can be used for the dimensional optimization design of groove geometry in both cylinder and face gas film seal.