Wood fiber-based insulating paper, valued for its renewability and electrical properties, is widely applied in oil-immersed power transformers. However, the high dielectric constants from hydroxyl groups in cellulose cause non-uniform electric field distributions, heightening insulation breakdown risk. This study employs molecular dynamics simulation to investigate the mechanism of direct fluorination in modulating the dielectric properties of cellulose insulating paper. The research is grounded in the principle of molecular dipole fluctuation, simulating the dipole autocorrelation functions (DACF) of cellulose molecules with varying degrees of fluorination. By applying the Kohlraush-Williams-Watts (KWW) expression to process DACF, complex dielectric constant spectra of different fluorination models were obtained and systematically analyzed. Fluorination reduced the complex permittivity and dielectric loss of cellulose molecules. In addition, the complex permittivity spectrum parameters of the cellulose fluoride model in the medium mixed with water were also investigated. Water makes the peak of the simulated dielectric spectrum move towards high frequency. By transcending the limitations of traditional experimental methods, this computational approach provides novel insights into molecular-level insulation paper modification, offering significant theoretical and engineering value for enhancing power transformer insulation performance.