Abstract: As a classical sound absorption coefficient prediction model, Delany-Bazley (DB) model has been widely used in sound absorption prediction of fiber materials. However, DB model can’t accurately predict the sound absorption performance of polyester fiber with special-sections. This study aims to optimize DB model for the sound absorption prediction of flat polyester fiber. The actual flow resistance data are measured to reassign parameter values in DB model. The optimized sound absorption prediction model is verified by the experimental data, and the error rate is analyzed. The results show that the optimized DB model can accurately predict the sound absorption coefficient of fiber materials with special-sections. The relative error between the predicted data and the experimental data decreases from 36.64% to 9.16%, the accuracy is improved by about 75.00%.