This paper considers estimation of transfer function in control systems. We will form discrete-type models of state equations and apply second order filter to its parameter estimation with the aid of the digital computer. Transfer function will first be transformed into state differential equations, which are converted to the discrete forms. Parameters in discrete-type state equations and inner real states constitute a new combined vector. We will adopt the second order filter to estimate the new vector involving unknown parameters because the resulting equation with the new vector is nonlinear. As numerical examples, estimation of the second ordere transfer function will be shown.