An iterative system identification scheme for the nonlinear vibration systems with random noise was presented. Firstly, for each sample of measured data with random disturbance, the derivative with respect to the unknown parameters was introduced into the cost function; then the difference equations for the unknown parameters were obtained by means of the discrete variational principle, and solved simultaneously with the revised constraint equations; thus, the tobeidentified parameters were guided from given initial conditions to their true numerical values with an iterative procedure. Secondly, for many data samples with random disturbance, the mean value of the tobeidentified parameters was obtained as the identification results. Finally, simulations for parameter identification of nonlinear vibration systems with fourdegreeoffreedom were presented so as to verify the proposed algorithm. Numerical results demonstrate that the parameter identification technology is effective.