The evaluation of the performance of decision-making units (DMUs) that use comparable inputs to produce related outputs can be accomplished through a non-parametric linear programming (LP) technique called Data Envelopment Analysis (DEA). However, the observed data are occasionally imprecise, ambiguous, inadequate, and inconsistent which may result in incorrect decision-making when these criteria are ignored. Neutrosophic Set (NS) is an extension of fuzzy sets which is used to represent unclear, erroneous, missing, and wrong information. This paper proposes a neutrosophic version of the DEA model, and a novel solution technique for Neutrosophic DEA (Neu-DEA) model. The possibility mean for triangular neutrosophic number (TNN) is redefined and modified the Khatter’s approach to convert directly the Neu-DEA model into its crisp DEA model. As a result, the Neu-DEA model is simplified to a crisp LP problem with a risk parameter (δ ∈ [0, 1]) that represents the attitude of the decision-maker towards taking risk. The efficiency score of the DMUs is computed by using various risk factors and divided into efficient and inefficient groups. The ranking of DMUs is determined by calculating the mean efficiency score of DMUs, which is based on various risk parameters. A numerical example is illustrated here to describe the suggested approach’s flexibility and authenticity and compared with some of the existing approaches.