To obtain intelligence information from uncertain environment in wartime, it will exist high threat factors for the HALE-UAV's flight safety because of long endurance, wide scope and large number objects of military reconnaissance and surveillance operation. To address such a situation, we develop a dynamic path planning method in this paper for UAV in-flight routing. We consider several problems using this method: (a) building a reference plan by the priori information; (b) consideration of two models, one is an environment information model based on the acquisition probability of targets, and the other is a threat model based on airborne sensors; (c) exploration of the decision system as a decision-maker which can permit UAVs to enter into threat zones with a known probability of being attacked in order to execute a reconnaissance mission; (d) determination the optimal reconnaissance flight path; and (e) evaluation the results of flight path planning.