Employing polytetrafluoroethylene (PTFE)-treated carbon fiber paper (CFP) as the substrate of the gas diffusion layer (GDL) is a common practice to improve water management in proton exchange membrane fuel cells (PEMFCs), but the resulting increase in electrical and thermal resistance is a critical problem that restricts the performance output of PEMFCs. Hence, studying the mechanism and prediction model for both the electrical and thermal conductivity in CFP is essential. This work established a mathematical graph theory model for CFP electrical and thermal conductivity prediction based on the observation and abstraction of the CFP characteristic structures. For the PTFE-treated CFP, the electrical and thermal conductivity of CFP can be effectively increased by optimizing the PTFE distribution in CFP. A "filter net effect" mechanism was proposed to reasonably explain PTFE distribution's influence on the CFP performance. Finally, the equivalent effect of multiple factors on conductivity was revealed using contour maps, which provides inspiration for further reducing the electrical and thermal resistance in CFP.