An innovative maintenance policy is proposed in this paper. This policy can involve spare parts ordering, production quality, and buffer inventory for an efficient production system. In the system, certain batches are required to be produced, and when each batch is finished, a determination is made whether maintenance is needed. The machine state deteriorates with the number of completed batches, and it can be improved by performing maintenance. Two types of maintenance activity, replacement and imperfect maintenance, can be selectively chosen to minimize cost. The defect rate of each batch is related to the number of completed production batches. An innovative concept, defined as the “virtual number” of completed production batches, is used to establish a link between maintenance and defect rate. Monte Carlo simulation and enumerative search is then used to determine cost-effective spare parts ordering and maintenance policies to minimize the cost for the production cycle. Finally, numerical examples are presented to demonstrate the model and to conduct sensitivity analysis. We find that in situations with a high buffer inventory costs, spare parts should be ordered late. When increasing the buffer inventory cost, more replacements should be performed compared to imperfect maintenance. Also, the buffer inventory cost rate and replacement duration time effect the rate of defective products significantly. These two parameters should be kept small, and controlled, if a very low defect rate is needed.