Crude oils can be classified according to their stability. The aromaticity of crude oil is one of the most important properties of oil samples. The aim of this dissertation was to investigate the correlation between the aromaticity of crude oil and the stability of crude oil samples. In this work, crude oils from different geographical origins were characterized using NMR spectroscopy and classified according to °API values. It has been shown that crude oils with a higher percentage of asphaltene have lower °API values. When investigating the correlation between the intrinsic stability of crude oil and the content of aromatic hydrogens (aromaticity) using 1H NMR, it was found that samples with lower crude oil aromaticity have higher intrinsic stability. The mass content of asphaltene and the diffusion coefficients were determined and used to estimate the hydrodynamic radii and the molecular masses. The results obtained by precipitation titration and spectroscopic methods were combined with statistical methods, multivariate linear regression and machine learning to develop a completely new model for predicting oxidation stability with an extremely good correlation factor for such mixtures.