Abstarct: Due to the fierce competition in the retail industry, the need to adopt an appropriate strategy in order to attract new customers and, more importantly, retain current customers, is felt more than ever. One of the tools that can help business managers in this direction is data mining tools. In this research, while introducing the nominal classification method with the selectability / rejectability approach, its performance in evaluating the customers' classification of a chain store has been evaluated. In this regard, using a simple sampling method, a sample of 200 was selected and the data collection process was performed by a questionnaire. The questionnaire questions, which are also considered as classification criteria, are designed baxsed on the SERVQUAL model. Also, due to the quality of some criteria and the possibility of doubt and hesitation in the answers of the respondents, hesitant fuzzy sets have been used to analyze the data. Finally, the results of data analysis showed that 20% of customers are classified as hateful customers, 15% as dissatisfied customers, 35% as regular customers and 30% as satisfied customers. Also, the final results of the research showed that the proposed method has significantly improved compared to similar methods.