Industrial robots are programmed to repeat a sequence of well-defined tasks, e.g. deburring, cutting and welding. The production cycle time is directly influenced by the task order as well as the way trajectories linking this tasks are generated. In this paper, we present an optimization approach that minimizes the production time as well as the overall movements duration of the robots. We propose a fast algorithm that generates a sequenced near-optimal solution for Multi- Robotic Task Sequencing Problem. We model the problem in the form of a new min (sum-max) Multiple Generalized Traveling Salesman Problem min(sum-max) MGTSP model. Near-optimal solutions are obtained to autonomously minimize cycle time in collision-free path. The originality of our method lies in its flexibility and ability to be integrated into industrial processes. In this study, we perform a double optimization in both the task and configuration space. Also, the proposed algorithm is able to automatically plan a collision-free trajectory between two robots configurations by generating relevant via points which minimizes movements duration. Comparing to other approaches, experiment reveal positive results in terms of efficiency and demonstrate the ability of this method to be integrated into existing industrial software.