Mathematical model for macromolecule catalytic conversion in a flow reactor includes four interconnected numerical calculations of different scales for the following phenomena: effect of increasing the concentration of coke grains and their size (nanometers, scale of coke particles) on porosity, tortuosity, and specific area of the catalyst computing percolation graphs of the mesoporous and hierarchically porous catalysts (dozens of nanometers, scale of percolation graph); kinetic patterns for asphaltene conversion and catalyst deactivation in the mesoporous and hierarchically porous pellets (millimeters, catalyst pellet scale); macrokinetic model for reactor operation filled with mesoporous and hierarchically porous pellets (centimeters, reactor scale). Mathematical instruments involves both discrete (Lubachevsky-Stillinger, Dijkstra algorithm) and continuous (Fick's law, kinetic equations) methods. Rate constants for kinetic modeling of the reactor operation were extracted by approximating the experimental points for the conversion of asphaltenes at the conditions close to industrial ones by numerically obtained curves. Striking difference in the texture evolution of mesoporous and hierarchical catalysts, observed by both catalytic experiments and theory, during asphaltene conversion (HDAs) resulted in fast deactivation of the first catalyst while the second one showed a long-term stability. The model is an excellent tool for the targeted design of high-performance hierarchical catalysts and catalytic layers and gives new possibilities in selection of the catalyst preparation ways.