A large database of full-scale plant experiment results was obtained in the steam cracking unit at Chemopetrol Litvinov Company. The results of full-scale experiments were used for the control and optimization of the cracker unit and for the kinetic model identification. In addition, the full-scale plant experiment results were used as the basis for laboratory testing using pyrolysis gas chromatography and to scale the laboratory experiment results up to the real process yields. Finally, the full-scale plant experiment results were used to develop an Artificial Neural Network model and characterize higher boiling feedstocks through a system of real hydrocarbon components.