To address insufficient stability in the bird-flocking-inspired phase transition control model for UAV swarms， an improved method based on phase transition rate is proposed. By comparing the behavior of natural bird flocks with that of the bird-flocking-inspired phase transition control model， the regulatory mechanism of the temporal effect of phase transition on group stability is revealed. Based on the leader-follower architecture， a control protocol including attraction， repulsion， alignment， and circling terms is constructed. The temporal characteristics are quantified by the dynamic adjustment rate of the coefficients of behavioral control terms， which is defined as the “phase transition rate”. Simulation experiments show that under the condition of a low phase transition rate， UAV swarms can achieve stable transitions from an ordered state to three behaviors： dispersion， aggregation， and circling. Comparative experiments indicate that the order parameter of low-rate phase transition is higher than 0.9， and the span of the stable interval is 3.67 times that of high-rate phase transition， which confirms that reducing the phase transition rate can improve the stability of behavioral transitions. Furthermore， a low phase transition rate inhibits the formation of vortex phases， indicating qualitative differences in behavioral patterns across different phase transition rates. This study demonstrates significant advantages in increasing the number of phase states in UAV swarms and improving phase transition stability. It not only verifies the key regulatory mechanism of phase transition rate on the phase transition process， but also provides new theoretical and methodological support for enhancing the engineering practicability of the bird-flocking-inspired model for UAV swarms.