Abstract:In the geodesy,many models are nonlinear.The classical methods dealing with nonlinear models are either linear approximations using the approximation value of parameters or iterations.Some nonlinear models can not be linearized and are very difficult to evaluate precision of estimators by the methods.To solve the problems,the penalized least square estimator for nonlinear models were presented and related estimation formulas supported.That is the key lying in that substituting the nonparameters of semiparametric models for the Taylor expansion reminders.The results calculated with the method show that the method is effective for the nonlinear models.