This work reports on a system for the Magnetic Imaging of Pyrolysis Feedstocks, acronym MIPF (pronounced with P silent), that was devised to model the yields of olefin products available from their cracking in a modern furnace. The system comprizes three facets, namely, sample preparation, NMR experiments, and data analysis. Sample preparation incorporates internal standards into the feedstocks to enable precise analysis of the NMR experiments, which provide quantitative C13 and H1 spectra, elaborated by 1- and 2-D procedures, such as DEPT and HETCOR. Data analysis employs (1) an Integral Regions train, which provides coarse but complete information about all the carbon and hydrogen atoms in a feedstock, particularly the aromatic C and H atoms, and (2) a Canonical Groups train, which provides high-level information about chemical moieties and compounds, but detects only about half of all the atoms in the feedstock, particularly those in n- and methyl-alkane chains. A preliminary “fingerprint” relation is presented between the MIPF parameters of five Naphtha feedstocks and their olefin and aromatic product yields from high severity steam cracking.