This paper is concerned with a novel approach to batch process automation using fuzzy modelling and reinforcement learning. The core part of the automation strategy is an autonomous agent that continuously learns to implement control actions that can drive the batch process’ state very close to the desired one with near-optimal performance. An efficient algorithm for reinforcement learning called fuzzy Q-learning is proposed to build the agent (controller). The use of linguistic information to guide the learning process and to implement near-optimal actions provides the means for both knowledge integration and scaling reinforcement learning. The methodology is exemplified using a batch process involving simultaneous reaction and distillation.