For spacecraft suffering from multiple celestial constraints in time-varying space enviroment, an on-board large angle attitude maneuver path planning based on particle swarm optimization was studied. The kinematics of the spacecraft was described by Rodrigues parameter. The full attitude was mapped to a point described in three-dimensional attitude space. Then the attitude path planning was transformed to the problem of three dimension path planning for point robots. The large angle attitude maneuver path planning algorithm based on particle swarm optimization for spacecraft with multiple celestial constraints was designed. It is necessary to change the attitude coordinate from the original reference frame to a proper reference frame which is more convenient in calculation for the path planning algorithm. Through the iterative program we can receive the feasible path. Numerical simulation results demonstrate the validity of the method.