Abstract: Considering the batch production format commonly seen in real-world manufacturing processes, a mathematical model for the batch scheduling problem in a permutation flowshop is developed, aiming to minimize the maximum completion time. To address this, an improved artificial rabbits optimization algorithm is proposed. During the encoding phase, the smallest position value (SPV) rule converts continuous solutions into discrete ones. In the decoding phase, a dynamic strategy is applied to group work pieces into batches. The NEH heuristic is employed to improve the quality of the initial population. Additionally, a differential evolution operator is introduced to enhance solution diversity. To further strengthen the algorithm's ability to avoid local optima, a local search strategy based on two-point exchange and inverse order is implemented. The performance of the proposed algorithm is validated through various test cases of different scales, involving fusion experiments, comparative analyses, and statistical tests. Finally, the algorithm is applied to solve a scheduling problem in the spraying workshop of an automotive exterior parts factory. The results demonstrate superior performance compared to other benchmark algorithms, further confirming the algorithm's effectiveness.