This study considers the set-up time between products. However, due to the difficulty of disassembly and assembly, the set-up time is different. Therefore, the asymmetrical set-up time and production batches are considered in the production scheduling system. Due to the practical needs of orders emergency insertion and dual-type customers, this study proposes a Random Optimization Algorithm (ROA) to search for the optimal scheduling under three performance indicators on demand and compare the results under two initial scheduling methods. The results showed that ROA is better than the initial scheduling methods. The random optimization algorithm proposed by this study can effectively improve the production performance.