Autonomous Unmanned Aerial Vehicles (UAVs) present a promising avenue for humanitarian missions, particularly in addressing urgent global public health concerns. Their versatility allows for the efficient transportation of lightweight payloads, including vital medical tests, blood products, vaccines, and medications , to remote, inaccessible, or perilous regions. In a military context, UAVs serve various critical functions such as surveillance, reconnaissance and logisti-cal support. The focus of this work is on Combat Service Support (CSS), a type of logistical support mission where UAVs provide supplies to ground forces personnel (GFs) engaged in combat or reconnaissance missions under threat. One challenge in CSS missions is the need for dynamic route planning, as threats can be discovered during the mission. Considering this, we propose a model named Safe Path Prioritization, which adopts as premise that the best way to avoid threat action is to prioritize regions where other UAVs have already flown and did not detect the presence of enemies. In addition, closer ground forces should have visitation priority, as the longer a UAV flies, the greater its chance of being neutralized. The proposed model is combined with a communication approach between UAVs and Ground Control Station (GCS) to increase the stealthiness of UAVs. We validated our proposal in diverse scenarios with different combinations of ground forces, number of UAVs and threat hostility level.Overall, the results show that the combination of safe path prioritization with a traditional minimum path approach can bring benefits regarding both the number of GFs visited and the number of UAVs kept unharmed in CSS scenarios.