Ensuring consistency and validity in flight training assessment proves challenging due to the traditional reliance on instructors' subjective evaluations and the imprecise quantitative criteria for competencies. This paper introduces a novel mathematical model based on the VENN model to evaluate behavioral indicators displayed during flight training across three dimensions: quantity, frequency, and the outcomes of threat and error management. This model maps these indicators to a quantified assessment of competencies, and historical checklist data is integrated to propose an optimization method for competency assessment criteria. Actual manual flight maneuver training serves as an example to verify the effectiveness of the proposed method. The results demonstrate a significant correlation between the final assessment criteria derived from the algorithm and the ratings provided by instructors, with 83.3% of the samples exhibiting rating consistency and the remaining samples deviating by no more than one level. This suggests that the proposed method adeptly transforms historical checklist information into assessment criteria for competency, offering theoretical support and a new data-driven approach for formulating competency assessment criteria.