Rough fuzzy sets (RFSs) offer significant advantages in handling uncertain information, but traditional models often rely on fixed equivalence relations, which limits their adaptability in dynamic environments. This paper introduces a novel dynamic rough fuzzy set ( \(\beta \) -DRFS) model, which incorporates fuzzy \(\beta \) -covering approximation spaces from an application-oriented perspective. In this framework, we define a fuzzy \(\beta \) -equivalence relation over a family of fuzzy sets, allowing the induced dynamic partition of the universe U varies with \(\beta \) . This dynamic adjustment enables the model to flexibly adapt to different levels of granularity in data analysis. Based on the dynamic equivalence relation, we construct a new rough fuzzy set model and systematically explore its mathematical properties. Furthermore, to enhance computational efficiency and logical precision, we develop a matrix method for calculating lower and upper approximations within the proposed model. Finally, we introduce a decision algorithm based on this model and demonstrate its effectiveness in controlling the Asian corn borer pest, showcasing its practical applicability in real-world uncertain data analysis.
