In this paper, the secure trajectory tracking issue is investigated for a quadrotor helicopter system with unknown-but-bounded (UBB) noises. It is assumed that the generated reference trajectory is sent to the helicopter via wireless networks which could be vulnerable to malicious attacks. The proposed encoding–decoding approach (EDA) can significantly improve the attack detection rate (DR), which is designed as a monotone and continuous function. Moreover, it is illustrated that both the DR and the tracking error can be affected by the parameter of the EDA. In addition, the zonotopic unknown input observer (ZUIO) is designed to estimate the helicopter’s states, which can be utilized to deal with UBB noises. Then, the gains of both the controller and the ZUIO can be calculated by minimizing the P−radius of the corresponding zonotopes. Finally, numerical examples are provided to demonstrate that the parameter of EDA can increase the DR and decrease the false alarm rate (FAR).