This paper presents a strategy to deal with the issue of air-to-ground trajectory planning for Unmanned Combat Aerial Vehicle (UCAV) considering multi-restriction by using a Radau Pseudospectral Method (RPM). First, the paper establishes and analyzes a 3-DOF UCAV model and various complicated constraint models such as the preliminary and terminal position, speed, attitude, flying performance, enemy fire or detection threats, the no-fly zone terrain and so on, then, an air-to-ground trajectory planning for UCAV considering multi-restriction model is founded. After analyzing the basic principle that RPM solves the optimal control problems and how to implement it, the RPM is adopted to transform the optimal control problem for trajectory planning into a nonlinear planning problem. Finally, the problem is solved by using the SNOPT software package, and a numeric simulation is carried out. The simulation results show that the RPM can generate a continuous, viable and optimal trajectory satisfying various complicated constraints.