To address the limitations of existing distance-based unmanned cluster collaborative navigation methods, which often overlook the impact of configuration on positioning capability and thus struggle to achieve precise navigation and positioning results, this study proposes a method based on genetic algorithms. The approach aims to minimize the sum of geometric dilution of precision (GDOP) values across all positions within a given task scenario, thereby determining the optimal anchor coordinates for different scenarios. Simulation results demonstrate that the navigation configuration based on GDOP minimization significantly enhances positioning accuracy and reduces positioning errors, providing effective support for the navigation of unmanned aerial vehicle clusters in satellite-denied environments.