The adaptive state tracking problem of switched systems is studied in this paper. The desirable state trajectory is generated by a switched reference model. First, a condition for asymptotical hyperstability of switched systems is proposed and a switching law is designed, which is a generalization of the classical hyperstability condition for non-switched systems. Then, the result is applied to uncertain switched systems to achieve state tracking. An individual adaptive law is designed for each subsystem such that the Popov inequality is satisfied. Asymptotical state tracking is achieved under non-persistent exciting input when the error system switches in a certain way. The result is demonstrated by a numerical example and a practical system of Highly Maneuverable Aircraft Technology vehicle, respectively. 29 and a few methods have been presented concerning the tracking problem [18][19][20][21]. As far as the model reference adaptive tracking problem is concerned, the direct adaptive control method using multiple Lyapunov functions plays an important role because multiple Lyapunov functions are less conservative in the stability analysis of switched systems. As the Lyapunov functions contain not only the state error but also the parameter estimation error, the values of the functions are not available, which make it difficult to test the condition of multiple Lyapunov functions method at the switching instant. To overcome this obstacle, the parameter adaptive laws are modified by a projection mechanism in [22,23]. Small tracking error is achieved by the dwell time method, and the tracking error converges to zero when the input signal is persistent exciting (PE). Asymptotical tracking is achieved in [24,25] for non-PE input, where the parameter estimation error in the Lyapunov functions is replaced by the parameter estimation so that the values of these functions can be obtained. In order to use the dwell time method, a variable structure term is added to ensure the decay rate of sub-Lyapunov functions. However, both methods demand the bounds of unknown parameters, which are usually unavailable in practice. So, it is necessary to find another method to remove this requirement in the designing procedure, which motivates us on this note.On the other hand, hyperstability is another powerful tool in designing the adaptive law for nonswitched uncertain systems in addition to the Lyapunov theory [26][27][28][29][30][31]. Hyperstability is a property of systems with bounded input energy [32]. It is derived from the dissipative theory and means that the motion of the system is bounded (and convergent) as long as the system is passive (strictly passive). As the parameter error only appears in the input of the error dynamics, there is no parameter error in the storage functions of the error systems. By using hyperstability theory, an adaptive law is designed to satisfy the Popov inequality, which restricts the total energy delivered to the error system. In this way, the problem of unavailable function values no longer exis...