The viscosity and rheological behavior of an ethylene glycol-water mixture based Fe3O4 nanofluid have been experimentally investigated. The nanofluids for this study were prepared by a two-step method in which Fe3O4 nanoparticles were added to a base fluid mixture consisting of 60% (w/w) ethylene glycol and 40% (w/w) water. The measurements were conducted at temperatures ranging from 288.15 to 343.15 K, and at nanoparticle volume fractions ranging from 0.0022 to 0.0055. Furthermore, the dependency of viscosity of nanofluids on shear rate was examined. The results indicate that increasing the shear rate leads to a reduction in the viscosity (shear thinning behavior). Finally, the obtained experimental data was correlated by both a thermodynamic model and a hybrid GMDH-type polynomial neural network, where the mean absolute relative deviation (MARD) of these models was calculated as 3.64% and 3.88%, respectively.