Abstract: In this study, in order to identify the mathematical model of an electric steering gear which is very useful for simulation and design of control system, an improved particle swarm optimization algorithm is proposed. Advantages and disadvantages of particle swarm algorithm are been investigated and demonstrated. The improved algorithm focus on the problem of stagnation in model identification, to solve the problem that particle swarm optimization is easy to fall into local optimal solution, the possibility methods are analyzed and a feasible algorithm is given. Through different discriminating methods, result of arithmetic operation is better than using particle swarm optimization only. With the improved algorithm identification, an experiment is done, the results show that the improved algorithm is very effective, can be used to identify dynamic model the improved particle swarm optimization algorithm can be used for solving the mathematical model of electric steering gear.