In the discrete manufacturing environment of a job-shop, an optimization method for the resource configuration adjustment in order to minimize its manufacturing and tardiness cost was presented. With the material flow and work flow of a job-shop, a manufacturing process simulation model was first constructed to evaluate the performance of each resource configuration adjustment properly. Then by combining the production simulation method and improved simulated annealing algorithm, a solution algorithm for adjusting the resource amount of work centers and the production routings of products (or the configuration structure and usage way of resources) jointly was designed. In the proposed algorithm, the search and selection of neighbor configuration alternatives was guided by the critical production path analysis, and its convergence rate was then accelerated. It can gain better solution quality and efficiency than the traditional random search method, thus is more suitable for modeling and solving the large-scale resource configuration problem.