Abstract: Combining the inverse system method with support vector machine,a multivariable nonlinear decoupling control strategy based on least squares support vector machine(LS-SVM) withαth order inverse system is proposed for the lysine fermentation process.Through analyzing the invertibility of this system,a nonlinear inverse model for lysine fermentation process is developed offline by using an LS-SVM with Gauss kernel.The obtained LS-SVM inverse system is connected in front of the original system,and the complicated nonlinear multivariable system is decoupled into several independent singleinput-single-output(SISO) pseudo-linear subsystems,so it can be controlled with the method of linear system.The simulation results demonstrate mat the LS-SVM inverse system decoupling control method has a good performance in approximating the nonlinear system.The method provides a new way to the decoupling control of multiple-input-multiple-output(MEMO) nonlinear systems.