<p>Traditional data envelopment analysis (DEA) models fail to adequately address uncertainty in inputs and outputs. In many real-world applications, uncertain quantitative and qualitative factors cannot be fully captured using a single traditional approach, such as stochastic optimization, robust optimization, or fuzzy programming, even when historical data are available. This study aims to develop a unified DEA-based framework capable of simultaneously addressing quantitative uncertainty from historical data and qualitative uncertainty expressed through linguistic evaluations. It integrates distributionally robust optimization (DRO) with interval-valued intuitionistic fuzzy sets (IVIFS) to capture different sources of uncertainty in a consistent framework. The DEA constraints are first categorized into quantitative and qualitative groups. For quantitative factors with historical data, we apply a Wasserstein metric-based DRO approach that does not require known probability distributions. For qualitative factors, we model the data using interval-valued intuitionistic fuzzy numbers and derive their corresponding deterministic equivalents. An important advantage of the proposed framework is that it does not require prior assumptions about probability distributions and can directly utilize observed data while preserving qualitative judgments. The resulting DRO-IVIFS-DEA model is nonlinear due to dual norm terms, but it can be reformulated as a linear program under norm-1 and norm-∞ settings, ensuring computational efficiency and practical applicability. A case study involving 35 suppliers of a home appliance manufacturer demonstrates the model’s effectiveness. Sensitivity analysis with respect to the Wasserstein ball radius confirms stable and robust performance under increasing uncertainty levels. Overall, the model offers a robust and flexible framework for performance evaluation in environments with complex, uncertain data.</p>