The revolutionary development and implementation of smart technologies have triggered the manufacturers’ servitization trend towards smart product service system (PSS). Accurate selection of smart product service (SPS) module is critical to successful planning and development of smart PSS concept. This study constructs a list of criteria for SPS module selection from the perspectives of service implementation, value symbiosis and smart capability. The selection can be deemed as a multi-criteria decision-making process including two parts: weight determination of criteria and module ranking, in which the intrapersonal linguistic ambiguousness and interpersonal preference randomness are involved. The best–worst method (BWM) is widely acknowledged as an efficient method for weight determination due to its superiority in quickly finding optimal weight with scant decision data. The data envelopment analysis (DEA) method is proven feasible to prioritize alternatives with cost-based and benefit-based criteria. However, these two methods cannot handle the uncertainties involved in the selection process which may lead to imprecise results. Moreover, the previous research rarely studies simultaneous handling of these two types of uncertainty in the realm of BWM and DEA. Therefore, the current study proposes a novel rough–fuzzy BWM-DEA approach to SPS module selection, with fully capturing both the intrapersonal and interpersonal uncertainties. The application of the proposed approach in the smart vehicle service module selection and the comparisons with other methods demonstrate the validity and effectiveness of the proposed approach.

智能技术的革命性发展和实施引发了制造商对智能产品服务系统（PSS）的服务化趋势。正确选择智​​能产品服务（SPS）模块对于成功规划和开发智能PSS概念至关重要。本研究从服务实施，价值共生和智能能力的角度构建了选择SPS模块的标准列表。选择可以看作是一个多标准决策过程，包括两个部分：标准权重确定和模块排名，其中涉及人际语言歧义和人际偏好随机性。最佳-最差方法（BWM）被公认为是一种有效的重量确定方法，因为它具有在决策数据少的情况下快速找到最佳重量的优势。事实证明，采用数据包络分析（DEA）方法可以优先选择基于成本和收益的标准。但是，这两种方法无法处理选择过程中涉及的不确定性，而不确定性可能导致结果不准确。而且，先前的研究很少研究在BWM和DEA领域同时处理这两种类型的不确定性。因此，当前的研究提出了一种新颖的粗模糊BWM-DEA方法来选择SPS模块，同时充分捕捉了人际和人际的不确定性。