Abstract: The paper presents a strategy for path planning of multi-mobile robots using genetic algorithms (GAs). Obstacles in the 2D workspace of the multi-mobile robots are simplified into convex polygons. The edges of the polygons are enlarged so that the size of the mobile robots can be scaled to points. MAKLINK Graph is used to model the workspace. Sequence numbers mark the cross points on the MAKLINK Graph. The sequence number is also used as a coding method for the chromosomes of GAs. Selecting randomly the cross points that connect the start point and the end one generates the initial population of the GAs. The length of the path is used as the corresponding fitness value of the chromosome. In order to avoid collision of any two mobile robots, a coordination matrix is put forward to decide the fitness values of the offspring. A simulation system is developed based on object-oriented method. Simulation results show that the method is effective in path planning of multi-mobile robots.