Modern production systems in modern industries are constructed with one or more support lines, such as rework, parallel, assembly line, and so on, to satisfy their objectives. Producing a rejected part is a common issue in manufacturing processes. Rework path (RP) is one of the most common types of production paths used to repair rejected parts and return them to the main line for reprocessing. The ideal buffer size in front of each machine tool bay in both branches of the production system should be optimized to create an efficient production system, including the main production line with rework path (MPL-RP). The optimization of buffer size of the system leads to the improvement in their overall performance, management, and operations. The key criteria influencing the buffer size decision and hence the MPL-RP production rate are the characteristics of machine tools, particularly their uptimes and downtimes. An efficient optimization model (EOM) is conceived in this study to determine the buffer size between each two neighboring machine tools in MPL-RP at any given uptimes and downtimes. To determine the ideal buffer size of the MPL and its RP, EOM employs a genetic algorithm (GA) based optimization model. The suggested method can estimate the ideal buffer size in an acceptable period of time, as demonstrated in the numerical examples, for small and large production lines, included in this study. As a result, when compared to previous selection approaches, the buffer sizes can be selected in a short span of time. The suggested approach can assist manufacturing engineers in making decisions during the design of new MPL-RP systems, as well as be utilized to improve the working of existing MPL-RP systems.