To compromise the gap between the linear and nonlinear controllers we made a study on how to design a linear time varying controller using virtual reference feedback tuning strategy.To identify the unknown time varying parameter vectors in the two linear time varying controllers we added a new regularization term about the two unknown time varying parameter vectors to the formal cost function.So we could constrain the time varying parameter vectors jump property in virtue of this regularization term.An iterative separable nonlinear least-squares identification method was proposed to estimate the unknown time varying parameter vectors.Through iterating the global optimum of the time varying parameter vectors could be obtained.By means of the convex optimization theory we derived a critical value and take it as the starting point for finding the appropriate regularization coefficient.Finally the simulation example results verified the validity of the identification method.