This paper presents a novel identification method for a dual-rate sampled data system with preload nonlinearity. By using a switching function, the nonlinear system is turned into an identification model. Then a missing output identification model based recursive least-squares algorithm is derived to identify the parameters of the system by all the inputs and outputs. Compared with the polynomial transformation technique, this method can estimate the unknown parameters directly and can decrease the number of the unknown parameters. The simulation results show the effectiveness of the proposed algorithm.