The PageRank model has been applied in sport ranking systems; however, prior implementations exhibited limitations and failed to produce valid rankings. This study analyzed 1,466 National Collegiate Athletic Association (NCAA) Division 1 football games and developed a novel, modified PageRank model. We also proposed an artificial intelligence-based ranking system that adapts the damping factor to match the specific characteristics of each dataset, optimizing it for each season. The modified PageRank model achieved a predictive validity of 71.6%, surpassing the performance of traditional PageRank (65.7%) and winning ratio (64.2%) methods. The implementation of this research holds the potential to enhance decision-making processes and provide valuable insights to stakeholders within the college football domain. Moreover, our modified PageRank-based rankings excel in discerning performance trends and patterns within conferences, enabling precise strategies for improvement and strategic planning, benefiting individual teams and entire conferences.