A method based on the set-membership identification theory for the parameter identification of a helicopter fully coupled flight dynamics model is presented. In view of the fact that the helicopter flight dynamics model is seriously coupled and that it is difficult to obtain explicit function relationships between the identification model and the parameters to be identified, an indirect identification algorithm is created in this method for the set-membership identification in the form of a state-space differential equation. The concept of generalized noise is introduced to implement the set-membership identification of a complex model which has numerous identification parameters that are seriously coupled and greatly differentiated in sensitivity. Then a two-step method for the set-membership identification of a helicopter fully coupled flight dynamics model is established. A comparison with the flight test data of helicopter BO-105 demonstrates that the identification model using this method has good coherence with the data. Compared with the traditional method which is based on statistic theory, the new method can enhance identification accuracy, speed and robustness.