Abstract: Solving the mixed-model scheduling problem is the most important goal for just-in-time production systems. The main desirable characteristics of ant algorithm are versatile, robust and parallel searching. This paper presents a novel co-operative agent approach, ant algorithm scheme, to solving the scheduling mixed-model assembly lines. The results show that the solution which ant algorithm produces is better than the one which Toyota's Goal Chasing Algorithm, Simulated annealing algorithm and GA produce.