A novel method of improved Dijkstra algorithm and particle swarm optimization is proposed to evaluate global path planning for mobile robot. The first step is to make the MAKLINK graph which is used to describe the working space of mobile robot. The limited length value of free linkage line is conducted to substitute the constant weights in the adjacent matrix, which is well correlated with the fact that the number of paths is drastically less than that using the conventional Dijkstra method. Then the particle swarm optimization is adopted to investigate the global path from the several possible paths. Therefore, the proposed method facilitates reducing the computing time which enhances the efficiency of particle swarm optimization when performs the global path planning for mobile robot. Furthermore, simulation result is provided to verify the effectiveness and practicability.