In order to solve the problem of bearing-only target location for single unmanned aerial vehicle in the dynamic battlefield environment, a new trajectory optimization algorithm based on dynamic window approach was proposed. The maximizing determinant of the Fisher information matrix was set as the trajectory criterion. In the dynamic battlefield, which consisted of dynamic detection radar and fixed/moving obstacles the criterion, based on the idea of dynamic window approach, was extended from the traditional single-step optimal principle to the evaluation of multi-step predicted trajectory. At the same time, the effects of the detection radar and fixed/moving obstacles on the predicted trajectory were also considered. The optimal heading of unmanned aerial vehicle was chosen by the moving horizon method. And the numerical simulations show that the algorithm enables the unmanned aerial vehicle to ensure the high-precision bearing-only target localization under the condition of effectively avoiding the radar threat and fixed/moving obstacles, which provides new ideas for solving the problem of single unmanned aerial vehicle bearing-only target localization in a dynamic battlefield environment.