A single-objective Flexible Job-Shop Scheduling Problem (FJSP) solving method based on the Reinforcement Learning-Remora Optimization Algorithm (RL-ROA) is proposed in this paper. The method integrates the characteristics of multi-variety small-batch production in enterprises and establishes a scheduling model of Multi-Variety Small-Batch Flexible Job-Shop Scheduling Problem (MVSB-FJSP) for minimizing the maximum completion time. Based on the reinforcement learning (RL) module, intelligent adjustments of the key parameters in the Remora Optimization Algorithm (ROA) are achieved, the algorithm’s computational efficiency and solution quality are enhanced. The effectiveness of the proposed method is validated through the Brandimarte standard examples, and comparisons are conducted with other algorithms. The results demonstrate that the RL-ROA exhibits higher reliability and superiority in solving the MVSB-FJSP.