Job-shop dynamic scheduling is an important subject in the fields of production management and combinatorial optimization. It is usually hard to achieve the optimal solution with classical methods due to the high computational complexity of the problem. A solution of job-shop scheduling problem based on multi-agent is presented for the comparability between the dynamic scheduling problem of job-shop production and the TSP problem. The dynamic scheduling of job-shop production is designed according to the pattern of TSP problem which can be applied with ACO. By the application case, the ACO is the new method to solve the dynamic scheduling of job-shop production.