Intravascular fibrin clots are resolved by plasmin acting at the interface of gel phasesubstrate and fluid-borne enzyme. The classic Michaelis.Menten kinetic scheme cannot describe satisfactorily this heterogeneous-phase proteolysis because it assumes homogeneous well-mixed conditions. A more suitable model for these spatial constraints,known as fractal kinetics, includes a time-dependence of the Michaelis coefficient Km(F) = Km0F (1+ t)h, where h is a fractal exponent of time, t. The aim of the present study was to build up and experimentally validate a mathematical model for surface-acting plasmin that can contribute to a better understanding of the factors that influence fibrinolytic rates. The kinetic model was fitted to turbidimetric data for fibrinolysis under various conditions. The model predicted Km0(F) = 1.98 μM and h = 0.25 for fibrin composed of thin fibers and Km0(F) = 5.01 μM and h = 0.16 for thick fibers in line with a slower macroscale lytic rate (due to a stronger clustering trend reflected in the h value) despite faster cleavage of individual thin fibers (seen as lower Km0(F) ). ε-Aminocaproic acid at 1 mM or 8 U/mL carboxypeptidase-B eliminated the time-dependence of Km F and increased the lysis rate suggesting a role of C-terminal lysines in the progressive clustering of plasmin. This fractal kinetic concept gained structural support from imaging techniques. Atomic force microscopy revealed significant changes in plasmin distribution on a patterned fibrinogen surface in line with the time-dependent clustering of fluorescent plasminogen in confocal laser microscopy. These data from complementary approaches support a mechanism for loss of plasmin activity resulting from C-terminal lysine-dependent redistribution of enzyme molecules on the fibrin surface.