Identification and evaluation of smart product service requirements (SPSRs) are considered to be the critical steps for smart product service system (smart PSS) designing. However, there are few studies on the SPSRs identification and lack of service evaluation methods that comprehensively consider multiple uncertain factors. Therefore, this paper proposes a framework for SPSRs identification, which innovatively considers the characteristics of smart PSS and helps to elicit SPSRs from the logic of data value creation. Furthermore, considering the effects of individual semantic vagueness and individual subjective preference, a method based on the rough-fuzzy number is proposed, which helps service designers obtain subjective and accurate evaluation results. The case of the smart maintenance service of smart cars is used to prove the flexibility and feasibility of the identification and evaluation method.