Aiming at the problems of slow response speed and unsatisfactory position control accuracy of traditional control methods in the position tracking control of target points in complex environments for forest firefighting unmanned aerial vehicles, an improved particle swarm optimization algorithm based on Bernoulli chaotic mapping initialization particles and dynamic adjustment of weights (BWPSO) was proposed to optimize the proportion integration differentiation (PID) control method. By using Bernoulli chaotic mapping and adaptive inertia weight strategy, the particle position update formula was improved, and a mathematical model of the control system was established and its transfer function was calculated. The controlled model was taken as the research object and simulated for verification. The results show that compared with the traditional PSO-PID algorithm, under the given target position and additional disturbance conditions, the regulation time of the BWPSO-PID algorithm is shortened by 76.3%, 72.8%(Disturbance 1), and 72.4%(Disturbance 2), respectively. It has a fast response and no overshoot. The steady-state error is reduced by 82.3%, 67.2%(Disturbance 1), and 75.5%(Disturbance 2). It has been proved that the dynamic response quality and robust performance of BWPSO-PID under different working conditions meet the compound control requirements of fast response and high precision in complex environments.