Aiming at the chaos dynamic behaviors existing in operation process of permanent magnet linear synchronous machine, the chaos model of machine was established. The largest Lyapunov exponents of the chaos model were calculated by programming with arbitrary parameter. Based on analyzing the chaotic dynamic characteristics of PMLSM, the state feedback decoupling controller was constructed and the order of the chaos system was reduced. The backstepping chaos controller was designed with state feedback decoupling model. According to uncertain parameters of PMLSM system, an adaptive backstepping method was proposed with decoupling model. The virtual control quantity were structured and control law were designed to real time track and forecast the system parameters. The stability analysis was carried out by constructing the Lyapunov equation to make the state of the system converge to zero quickly and stably. The simulation results show that the PMLSM chaotic system could quickly restore to a stable operation status with the proposed decoupling adaptive backstepping method with strong robustness, fast response and high control accuracy.