开发了一种基于碰撞罚函数的移动机器人路径规划神经网络算法,充分利用神经网络的融合性和并行性提高移动机器人路径规划算法的运算速度,取得了显著效果.并提出了局部路径修正和路径“张紧”两种算法,有效的解决了局部路径微,J、碰撞和最短路径问题,实现了机器人路径的最优化.最后给出了仿真结果.“,”This psper develops a neural network algorthm of mobile robot\'s path planning based on collision penalty function, it makes great advantage of properties of its parallel process and data fusion of neuralnetwork to improve calculating speed of the algorithm and get an striking effect. At the same time we propsedtwo algorithms: revising path partially and path tension respectively, it solved local small collision and shortestpath problem and fulfill the optimality of robot\'s path. At last an emulating experiment is given.