Attitude control of an unstable launch vehicle (ULV) is crucial for maintaining an accurate spatial orientation and ensuring a precise location for communication, maneuvering, and scientific research. Classical proportional-integral-derivative (PID) controllers are frequently used to control spacecraft attitude owing to their simplicity. However, their robustness and agility in response to dynamic space conditions and disturbances do not meet the required design standards. To improve the system stability, in this study, we propose a fractional-order PID (FOPID) controller with a setpoint filter, which is optimally tuned using the zebra optimization algorithm (ZOA). The performance of the proposed FOPID–ZOA controller is compared with that of a traditional PID controller. The proposed optimization strategy is applied offline to tune the controller parameters, which remain constant during simulation tests. The simulation results show that the proposed approach outperforms the classical PID in terms of robustness, accuracy, disturbance rejection, and overall performance. Notably, compared with the traditional PID controller, the proposed controller requires a significantly lower control voltage during the transient response. Thus, the peak voltage is reduced from 20 V to 1 V (or approximately 95% less), which in turn reduces energy consumption and improves system stability. Therefore, the proposed approach represents a promising solution for systems exhibiting ULV-like dynamics, particularly where energy efficiency is a primary concern.