The clustering of copper flotation process has a significant impact on the performance of operation adjustment. Nowadays, due to the complexity of the copper flotation process, the adjustment of operational variables, which are controlled by operators, is often not regulated properly in time. Therefore, it is necessary to obtain a clustering strategy for the copper flotation process to guide the operators by taking prompt and effective adjustment strategies. Due to the uncertainty of clustering itself, the number of categories and their respective probabilities are needed. Based on affinity propagation (AP) clustering algorithm and gaussian mixture model (GMM), a clustering algorithm is proposed in this paper, which is referred to as AP-GMM. It can get the optimal number of categories and their respective probabilities without giving the number of categories in advance. On this basis, a new category matching theory is constructed. Inspired by rewards and punishments in reinforcement learning, a penalty function is investigated to verify the optimum number of condition categories for copper flotation process. Finally, experiments show the effectiveness and feasibility of the proposed method.