The main purpose of this work is to model dynamically the latero-torsional response of the structure of a landing gear using identification techniques. The emphasis is placed on identifying this motion on aircraft twin-wheeled telescopic main landing gears. Between the many possible model structures, a linear polynomial ARX and ARMAX model structures and a state space model structure were adopted, all of them discrete models. Structures of the Output-Error (OE) e Box-Jenkins (BJ) type were also analysed but were discarted. The coefficients of the ARX model are obtained through least square techniques and the ones for the ARMAX by using the extended least square estimator. The ones for the state space model are obtained through subspace projection techniques for obtaining the states e also by least square for obtaining the dynamic matrices. The system dynamical equations are developed for a better understanding of the physical problem. The landing gear lateral and torsional structural deflections, wheels rotations and tire angular slip are considered as degrees of freedom of the model. The problem physics and correlation analysis was used for obtaining the polynomial model and state space model order and input delay respectively. For the ARX and ARMAX models, it is proposed an order reduction and obtention of a second order model. Using the transfer function obtained from this model it is possible to find the frequency and damping of the landing gear modes. For the state space model, through the obtention of the dynamical matrix and its eigenvalues it is possible to find the frequency and damping of the landing gear modes. The results obtained through these techniques are compared.