The intention of the target can help us to estimate its future motion state more accurately. This paper proposes an intention-aware planner to enhance safety and robustness in aerial tracking applications. Firstly, we utilize the Mediapipe framework to estimate target's pose. A risk assessment function and a state observation function are designed to predict the target intention. Afterwards, an intention-driven hybrid A* method is proposed for target motion prediction, ensuring that the target's future positions align with its intention. Finally, an intention-aware optimization approach, in conjunction with particular penalty formulations, is designed to generate a spatial-temporal optimal trajectory. Benchmark comparisons validate the superior performance of our proposed methodology across diverse scenarios. This is attributed to the integration of the target intention into the planner through coupled formulations.