This study aims to evaluate students’ cognitive attribute mastery across several mathematics domains using the Cognitive Diagnostic Model (CDM) approach. The study involved 219 eighth-grade students who were assessed in five mathematics domains: basic statistics, sequences and series, relations and functions, linear equations and gradients, and systems of linear equations and inequalities. The analysis was conducted to identify patterns of students’ skill mastery and to determine the most appropriate diagnostic model for each domain. The results indicate that each mathematics domain exhibits different cognitive structures. In the domains of basic statistics and sequences–series, attribute mastery tends to be more flexible and partially compensatory, although in basic statistics there is evidence that conceptual understanding of statistical concepts functions as a prerequisite for calculating measures of central tendency and dispersion. In contrast, in the domains of relations and functions, linear equations and gradients, and systems of linear equations and inequalities, the pattern of attribute mastery appears more hierarchical and non-compensatory, where mastery of fundamental concepts becomes a foundation before students can perform procedural calculations and apply the concepts. These findings suggest that evaluating mathematical ability solely through total test scores is insufficient, as it does not capture the specific learning difficulties experienced by students. Therefore, the application of CDM provides more detailed diagnostic information regarding students’ strengths and weaknesses at the attribute level, which can serve as a basis for designing more targeted and effective instructional strategies.