<p>This paper introduces a novel generalization of N-Soft Sets, referred to as R-Soft Sets, which allow for real-valued ratings rather than only nonnegative integers. R-Soft Sets are particularly suitable for representing uncertainty in evaluation and decision-making scenarios where continuous assessments are needed. We formally define this structure and investigate its algebraic properties, including intersection, union, and various types of complement operations. A new entropy measure tailored for R-Soft Sets is proposed to quantify the level of hesitation in assigning non-integer values, particularly those close to midpoint values within a scoring interval. This measure generalizes entropy concepts from existing Soft Set and Fuzzy Soft Set theories. We prove that the proposed function satisfies the required conditions of a valid entropy measure. Furthermore, two decision-making algorithms are developed based on the R-Soft Set framework, which incorporate both unweighted and weighted evaluations. Examples and tabular representations are provided to demonstrate the applicability and effectiveness of the proposed models in supporting decision analysis.</p>