Abstract: A virtual obstacles based path planning method for mobile robot in urban road environment is presented.It's based on both the benefits of global and local path planners.Firstly,it uses A^* algorithm to produce a global optimal path.Secondly,virtual obstacles are generated according to the global path.Finally,by fusing virtual and actual detected obstacles,the local path is planned based on the improved vector field histogram method.This method can not only fully utilize environment information to get the global optimal path,but also avoid the stochastic obstacles on the road in real-time. Practical experiments illustrate the effectiveness and robustness of this method.