Abstract: The concept of universal shop scheduling problem (USP) was proposed based on the analysis of a real production system. A hybrid ant colony optimization (HACO) was developed to be applied to the USP. The results were compared with those of genetic algorithm. The numerical experiments make use of several benchmark instances whose scale is up to 75×20. Minimizing makespan is taken as the objective function. Both termination conditions of computation generation and solution convergence are tested for the computation. From the numerical experiments, it can be seen that when the computation generation is kept the same, HACO will make the solutions more convergent, and when the convergency is kept the same, HACO will consume less time.