<p>The dynamic consensus of multi-agent systems (MASs) comprises a significant area of contemporary research. This study systematically investigates and compares the effectiveness of three prominent sliding mode control techniques: conventional sliding mode control (SMC), twisting second-order sliding mode control (Twisting SMC), and super-twisting sliding mode control (STSMC) in enabling dynamic consensus among multi-agent unmanned aerial vehicle (UAV) systems. The evaluation of these controllers is executed within a unified UAV consensus framework, employing graph theory and a Laplacian-based topology representation to model the interactions among agents. This study evaluates performance metrics such as tracking accuracy, speed of convergence, and the ability to suppress chattering. To build upon the insights acquired, a fractional-order super-twisting sliding mode control (FO-STSMC) strategy with short-memory principle is proposed, which aims to enhance disturbance rejection and strengthen transient response, even in the case of packet loss and communication delay. The introduction of fractional-order dynamics incorporates memory effects into the control law, thereby enabling more precise tuning of convergence dynamics and reducing overall control effort. Simulation results point out that the proposed FO-STSMC remarkably outperforms the performance of integer-order STSMC with respect to tracking precision and chattering suppression, even in the presence of substantial external disturbances.</p>