This paper presents an adaptive methodology for optimizing dwell time allocation in group target tracking. The approach formulates dwell time allocation as an optimization problem characterized by a cost function that integrates resource utilization and tracking accuracy. Enhancing the quality of observations is essential for improving tracking accuracy in the geometry and kinematics of group targets. This quality is influenced by the covariance of group target observations and the number of detected targets. This research clarifies the relationship between the covariance of group target observations and the number of detected targets concerning dwell time. Additionally, an analytical relationship is established between the Bayesian Cramér-Rao lower bound (BCRLB) and dwell time for evaluating accuracy in the geometry and kinematics of group target tracking. The proposed adaptive strategy is recursively optimized throughout the tracking cycle, demonstrating significant improvements over conventional methods. Simulation results indicate that optimal dwell time allocation can achieve a fourfold reduction in error compared to a constant dwell time approach, as well as a fivefold decrease in resource consumption relative to conventional methods that either detect all group members or employ a fixed dwell time.