This work aims to predict the noise reduction coefficient (NRC) of nonwoven layers produced from Estabragh/polypropylene blends of fibers. Estabragh is a natural hollow fiber which can be expected to have excellent sound insulation properties. In addition to the investigation of Estabragh’s unique properties in association with the sound absorption capability, the effects of some predictor variables such as blend ratio, punch density, and areal density (g/m²) on the acoustical performance of the produced nonwoven samples also have been studied through this research. For this achievement, a simplex lattice design was used in which the mixture-process crossed regression models with two mixture components and two process variables (punch density and areal density) were built to predict the NRC values of samples. Statistical analysis steps were implemented using Design-Expert statistical software. The correlation coefficient between the predicted and measured values of NRC was found to be 0.98, which confirmed the validity of the developed regression model. It was found that the noise absorption behavior of samples increases as the proportion of Estabragh fibers in the blend increases. Additionally, results showed that samples subjected to more severe needling operation exhibit higher values of NRC. It was also observed that the areal density of samples directly affects their acoustical performance. The higher areal density leads to more sound absorption ability of samples.