In order to improve the field efficiency and reduce the operation cost, a research was conducted on the path planning method for the sprayer UAV in the field with 3D terrain. Firstly, aiming at building a 3D environment model, the grid method was selected to divide the field into small grids with the initialized properties. Secondly, the UAV was ordered to move from the current grid to the adjacent one with the highest probability. And therefore a coverage path which moved from one extreme of the field to the other in direction parallel to the crop rows alternately and turned at the boundary was identified. Thirdly, a mathematical model was established, of which the objective was to obtain the optimal return points with the minimum time in the non-spraying mode. Once the gravitational search algorithm was applied to solve the model, the planned path with return points would be outputted automatically by the proposed method. Furthermore, the proposed method was compared with the method for 2D terrain in a 700m×100m field, of which the result showed that there was deviation between the positions of the return points calculated by the different methods. And the same field was taken to test the performance of the proposed method, in which the proposed method was compared with simple path plan method and unplanned path respectively. Results showed that the advantages of the proposed method at the distance of the round trips was reduced by 23% and 90%, while the non-spraying time was reduced by 7% and 54%, respectively. After that, the proposed method was applied to a real field with 3D terrain. And the distance of the round trips and the non-spraying time were reduced by 11% and 5%, respectively in the experiment. Finally, the research result indicated that the proposed method was a reasonable, feasible and useful alternative to produce paths with less time for the sprayer UAV.