Requirements elicitation is a key sub-process of requirements engineering in which various types of software requirements such as functional and non-functional requirements, testing requirements etc., are identified according to the needs of the stakeholders. After the completion of the requirements elicitation process, there may be a large set of requirements and practically it is not possible to implement all requirements due to budget, time and other constraints of an organization. Under this situation, software developers select the highest ranked software requirements from the list of the elicited software requirements for different releases of software so that a successful software product can be developed. The selection of software requirements for different releases is a difficult task because it requires the participation of various stakeholders. In the field of science and engineering, datasets play an important role during experimental work. Based on our review, we found that in the literature of software engineering less attention is given on the generation of the datasets particularly in the domain of software requirements prioritization and selection (SRPS). Therefore, to address this issue, in this paper we present a four-step method for generating the dataset for SRPS research, i.e. (a) identification of stakeholders, (b) elicitation of software requirements using the goal-oriented concept, (c) formation of decision maker committee and (d) evaluation of the functional requirements and non-functional requirements by decision makers under fuzzy environment. The proposed method has been applied to generate the dataset for the requirements of an institute examination system.