The relaxation method is well-known approach to stabilizing a nonlinear simulation procedure. Relaxing the solution may slow down the convergence rate but will make the nonlinear iteration more stable. It is helpful for the highly nonlinear-coupled physics problems.  In this talk, we introduce our most recent result in this method for a class of nonlinear prarabolic problems. The method is controlled by a relaxation parameter and the optimal choice of such a parameter is the maximum value that achieves consistent convergence behavior of the nonlinear system, although it is not easy to obtain in practice. We will show details on how to fix this parameter.