With the gradual opening of the low-altitude airspace and the rapid development of Unmanned Aerial Vehicle(UAV) industry, the users of UAV are increasing continuously and the conflicts could occur at any time. It is necessary to develop a reliable UAV conflict resolution algorithm to avoid the danger. This paper proposes an UAV conflict resolution algorithm based on the improved ant colony algorithm with two advantages. Firstly, the algorithm adopts adaptive parameters adjustment strategy, which adjusts the parameters value dynamically according to the quality of the solution, prevents the algorithm premature convergence and improves the accuracy. In addition, the disturbance factors is introduced to the state transition rules of random selected path in order to accelerate the initial convergence. The simulation results have shown that the improved algorithm displays obvious superiority in convergence precision, helping the two UAVs avoiding dangers in time. The algorithm described in this paper could be applied to target identification, path planning and other issues as a general optimized algorithm, which is of great significance and wide application.