Abstract: Dynamic soaring technology is one of the critical approaches to improving the endurance performance of unmanned aerial vehicles (UAVs). Nevertheless, existing studies neglect the impacts of high-altitude sideslip, resulting in poor adaptability of trajectory planning frameworks and low efficiency in online solution, which makes it difficult to balance accuracy and real-time performance.Aiming at high-altitude environments with strong crosswinds, this paper established a sideslip-corrected dynamic model by introducing sideslip effects and thrust compensation to accurately characterize the real force characteristics of UAVs under high-altitude wind fields. A three-stage hierarchical planning framework consisting of offline pre-planning, online reference trajectory generation, and real-time correction was designed to adapt to dynamic wind-field variations and address the challenge of rapid online trajectory planning. Furthermore, a segmented sequential convex optimization method was proposed, which segments trajectories according to dynamic soaring features and transforms them into quadratic programming problems for fast iterative solution. Simulation results show that compared with the traditional Gauss pseudospectral method, the single-run solution time of the proposed method is reduced from 130.3 s to 0.88 s. In contrast to the constant-altitude non-soaring trajectory, the energy output is decreased by 17.9%, which significantly improves the energy utilization efficiency of UAVs.