Identification and characterization of asphaltenes, the most polar and predominantly aromatic components of c-rude oil, still post a challenge for researchers in the oil industry, owing to their complex molecular architecture. Recently, much effort has been devoted to understanding their structure and physicochemical properties. In this paper, a combination of diffusion-ordered NMR spectroscopy (DOSY) and statistical multiway methods was used to investigate petroleum samples and asphaltenes of different origin. Such multiway TUCKER3 decomposition has been used for the first time to. analyze the two-dimensional matrix of complex nuclear magnetic resonance (NMR) data for petroleum samples. DOSY NMR. spectra of 50 samples (45 for establishing the statistical model and 5 for its validation) were recorded and evaluated by an in-house-developed code incorporating multiway analysis to set up a tool for predicting their identity and origin. A statistical model was developed E that can:identify and separate asphaltene samples from those of crude oils, vacuum and atmospheric residues, and resins which is not possible directly from the NMR spectra. Furthermore, the model clearly demonstrated that all asphaltene: samples clustered. into the same group, irrespective of the geographical origin, implicating their structural and size similarities. The validity of the model was further tested by analyzing an additional set of asphaltenes. It has been demonstrated that the proposed approach is very useful for analyzing complex oil mixtures, and has the potential for developing a robust quantitative model.