The methods of electronic throttle experimental identification together with optimization based parameters estimation are presented in this article . The parameters identified by this experimental identification were used as initial values of optimization variables in the problem of nonlinear optimization, where the cost function was non smooth. For this type of problem the Pattern Search method was used and satisfactory results were obtained. The results in this article are presented in the graphical form and show closeness between the measured and modeled signals of throttle plate position and angular velocity. The results are satisfactory for data, which were measured in closed loop by nonlinear throttle plate position controller.