To resolve the conflict between precision and workload in path planning of an UAV with artificial fish-swarm algorithm based on gridding method, an improved artificial fish-swarm algorithm was presented. The proposed algorithm introduces adaptive step length method and adaptive gridding segmented traversal method. In early iteration, the proposed algorithm calculates with large step length, later calculates with narrowed step length and gridding segmented traversal method to find a better solution around a defective solution. The simulation result shows that the proposed artificial fish-swarm algorithm is more accurate and more stable than basic artificial fish-swarm algorithm or adaptive step length artificial fish-swarm algorithm and has large advantage on workload over artificial fish-swarm algorithm based on gridding method.