To provide a potential method for predicting friction coefficients with pavement texture parameters, the Hilbert-Huang Transformation was used to extract relevant parameters from the pavement profile. Eight parameters of the Optimal Intrinsic Mode Function (OIMF) were calculated based on the Hilbert-Huang Transformation. A correlation analysis was conducted between these parameters and friction, taking into account the influence of velocity. Furthermore, a multiple linear regression model was developed to explore friction prediction. The results revealed that the parameter representing the peak density divided by sharpness in the OIMF exhibited the strongest correlation with friction (R2 =0.78), with this correlation being influenced by velocity. The multiple regression model may provide a reliable method for friction prediction (R2 =0.77), and the predicted friction coefficients demonstrated no significant deviation from the measured values (p-value < 0.0001).