Using a rotary viscometer Brookfield DV-II+Pro, the viscosity of an almost one-component (1-2% impurity) sample of synthanol ALM-7 was studied. In the presented work, this reagent is use as a sample of a highly viscous non-Newtonian fluid and a concentrated micellar disperse system, the particles of the dispersed phase in which are micelles from molecules of this surfactant with dimensions less than 10 nm. Using the example of such a fluid, it is shown that the decrease in viscosity observed in it, typical for dispersed systems, as the shear rate increases, is accompanied by an increase in the activation energy of the viscous flow, which is inconsistent with the Arrhenius and Frenkel equation. The reason is that these equations do not take into account the changes in entropy ∆S during the viscous flow of the non-Newtonian fluid, the value of which actually determines the sign of the change in the viscosity of the non-Newtonian fluid with increasing velocity or shear stress. The only way to calculate ∆S now based on the use of the Eyring equation. However, for the correct calculation of ∆S by the temperature dependence of the dynamic viscosity of the non-Newtonian fluid and the Eyring equation, an independent correct way of finding the value of the preexponent B in this equation is necessary. The article analyzes the methods described in the literature for calculating the values of B, including those proposed by Henry Eyring himself. As a result, it was revealed that only the experimental method we developed for estimating the values of B corresponds to real processes in the non-Newtonian fluid, since only with such calculations does an increase in temperature and shear deformations lead to values of ∆S > 0, indicating the destructive effect of these factors on the non-Newtonian fluid. It is shown that other methods of calculating B can lead to incorrect values of ∆S < 0 and, as a consequence, erroneous conclusions about the processes occurring inside the non-Newtonian fluid.