Sustainable sludge management is becoming a major issue for wastewater treatment plants due to increasing urban populations and tightening environmental regulations for conventional sludge disposal methods. To address this problem, a good understanding of sludge behavior is vital to improve and optimize the current state of wastewater treatment operations. Rheology being the science that studies the deformation and flow of matter, this behavior is characterized from the sludge rheological properties. Up to now, performing these measurements is long and complex. The aim of this study is to assess the potential of Polarized Light Spectroscopy to predict the sludge’s main rheological parameter. The challenge is big as (i) sludge are very complex material (opaque and very turbid) from which it is difficult to measure a signal of high quality and as (ii) NIRS is mostly used to extract chemically related information from materials, while in this study, the objective is to predict some properties highly related to the physical structure. To overcome these issues, a specific optical setup, implementing light polarization, allowed to filter different components of the diffusely reflected light: part of the signal that has properties close to ballistic photons (i.e low scattered photons) and part of the signal that has been highly scattered, and which carries information about the materials structure. A various panel of 34 sludge samples (primary, secondary, digested, and dehydrated) was collected in different wastewatertreatment plants in France. The multiscattered signal (RMS) has been used to predict, with Partial Least Square, some of the most relevant rheological parameter: viscosity, yield stress, elastic and viscous modulii. Compared to the models built with a classical reflectance (RT), RMS offers better figures of merit (table 1). On the contrary, to predict the sludge’s dry matter, it appeared to be more relevant to use the low scattered signal (RSS), as dry matter is closely dependent to the chemical properties of the material. This study shows that even on highly opaque and turbid material such as sludge, it is possible, by using wave properties of light, to select part of the NIR signal presenting a higher correlation to physical structure of the sludge and being useful to predict sludge's rheological properties.