<p>Programming education consistently faces challenges in bridging theory with practice and fostering students’ cognitive competencies. This 12-year longitudinal study (2011–2023) investigates an innovative competency-based teaching model in university C programming education that integrates six educational theories into a coherent framework with three dimensions (theoretical, practical, innovative), four integration mechanisms, and five combinatorial strategies. Using a mixed-methods approach with a quasi-experimental design, we studied 4,051 undergraduate students from a Chinese university. Results revealed significant enhancement in students’ cognitive abilities, as measured by Raven’s Standard Progressive Matrices (<i>t</i> (350) = 8.76, <i>p</i> &lt; 0.001, <i>d</i> = 0.68), which strongly correlated with improved academic performance (<i>r</i> = 0.62), computational thinking (<i>r</i> = 0.71), and problem-solving skills (<i>r</i> = 0.67). The model creates multiple pathways for cognitive development through synergistic interactions between components, promoting collaboration and self-directed learning with effects extending beyond graduation. Multiple regression analysis identified three key predictors of cognitive enhancement: classroom engagement (<i>β</i> = 0.35), project completion (<i>β</i> = 0.28), and participation in innovation activities (<i>β</i> = 0.22). This study provides robust empirical evidence for the long-term efficacy of a competency-based model in programming education, presenting a transformative approach to STEM education reform particularly relevant in rapidly evolving technological landscapes.</p>