This study addresses the challenges of subjectivity and inconsistency in expert-based wood quality assessment by proposing a novel hybrid model. The model integrates Grey Relational Analysis (GRA) and the Kappa coefficient to provide a systematic, two-dimensional evaluation, simultaneously quantifying expert decision accuracy and consistency. The framework is applied to industrial data from ten experts assessing 28 wood quality attributes. Categorical grades are processed via numerical encoding and a grey possibility function. A key analysis compares the reliability of GRA and the weighted Kappa coefficient under small-sample conditions (10–20 samples). Results show that while both methods produce a coherent expert ranking, GRA exhibits superior stability and lower variability with limited data, suggesting greater robustness in such scenarios. The model also identifies specific weaknesses in the evaluation process, particularly for borderline qualification grades, providing actionable insights for training. The proposed tool offers a practical solution for expert evaluation in data-scarce environments. Its main limitation is the dependence on a subjective expert reference standard, indicating a need for validation with objective benchmarks and larger datasets.