In recent times, nanofluids have been investigated comprehensively by the scientists because of the progresses in nanotechnology. Nanofluids' physical properties are necessary to be known accurately in order to reduce the error rate of models. Estimation of viscosity value of nanofluids is a significant one owing to its extensive usage in thermal engineering. In this work, carbon nanotube (CNT) particles are chosen to prepare nanofluid blend with its pure water and antifreeze. Their volumetric concentrations are varied from 0 to 2%in both pure water and antifreeze having various proportions with pure water. As soon as nanofluid having some surfactants becomes stabilized by means a sonicator and ultrasonic bath, viscosity meter is used to determine the viscosity regarding with the temperatures ranging from 15 degrees C to 50 degrees C. Consequently, the measured viscosity values decreases with nanofluid's temperature and increase with particle concentration as expectedly. In addition, a comparison study with the measured results and almost all empirical correlations in open sources has been done in order to determine the most predictive ones for the studied nanofluids. It was derived from the results that viscosity correlations can estimate different sorts of nanofluids' viscosity even though their base fluids with particles having various diameters are totally different from others. (C) 2016 Elsevier Ltd. All rights reserved.