Styrene-butadiene rubber (SBR) is a major source material for the fabrication of elastomer products. Depending on its origin, differences are observed between SBR samples by tabletop NMR spectroscopy that relate to the constitution of the macromolecular chains. This study reports experimental results from the analysis of 108 SBR samples by low-field H-1 and C-13 NMR spectroscopy at 1 T in combination with partial least squares regression to develop a methodology for quality control of raw rubber. Partial least squares regression (PLS-R) models were developed for quantifying the individual monomer units present in SBR which are impossible to quantify directly because of peak overlap in a 1H NMR spectrum obtained at 1 T. The spectra revealed differences between samples from the same and different manufacturing batches of the same and different manufacturers in a qualitative and quantitative fashion. The range of samples included regular and oil-extended solution and emulsion polymerized SBR. Referring to high-field spectra acquired at 9.4 T the peaks in the low-field C-13 NMR spectra could be assigned for determining the rubber microstructure, and the content of different repeat units could be quantified by partial least squares regression. Over 12 repeatable measurements the standard deviation in mass % was 0.03%, 0.06%, 0.05%, 0.33% and 0.37% for the contents of styrene, 1,2-butadiene, 1,4-butadiene, trans-1,4-butadiene and cis-1,4-butadiene units, respectively. Among 7 different sampling points in a delivery, the standard deviation was 0.51% for 1,2-butadiene, 0.88% for styrene, 0.56% for 1,4-butadiene unit, 0.42% for trans-1,4-butadiene and 0.68% for cis-1,4-butadiene units. The root-meansquare error of prediction (RMSEP) for styrene, 1,2-butadiene, 1,4-butadiene, and trans-1,4-butadiene was 0.15, 0.29, 0.29, and 0.28 with R-2 values of 0.93, 0.92, 0.92, and 0.95, respectively, demonstrating the potential of low-field NMR spectroscopy with compact instruments for quality control of raw rubber when used in combination with effective data analysis procedures such as chemometrics. (c) 2018 Elsevier Ltd. All rights reserved.