Heavy crude oils are commonly mixed with lighter crude oils to reduce heavy oil viscosity and to ensure the flow through the pipeline system. Viscosity, as is well known, is a crucial property for designing and sizing offshore and onshore facilities for producing and transporting crude oil. When not available experimentally, the viscosity of mixed crudes can be obtained from theoretical approaches such as the Arrhenius-Lederer equation depending only on bulk properties (density and viscosity) of mixture components, and an adjustable parameter (α), which allows calculations to match experimental data. This work presents a new correlation for calculating the rescaled parameter proposed by Lederer for the Arrhenius equation to determine the viscosity mixtures of heavy and light hydrocarbons. The proposed correlation was developed considering experimental data of binary mixtures of paraffinic hydrocarbons, diesel, and heavy oil over a wide range of temperature conditions (from 293 to 343 K), fluid density ratios (from 1.014 to 1.40), and viscosity ratios (1.3 to 311,000). For viscous fluids, which exhibited non-Newtonian behavior, viscosity on the first Newtonian region is considered. The correlation was tested to predict the viscosity of biodiesel-diesel, heavy crude oil-diesel, bitumen-heptane, and eicosane-decane binary mixtures from 308 to 448 K and pressures up to 10 MPa. Finally, for assessing the predictive capabilities of the new correlation, the viscosity of several mixtures was estimated. The first consisted of mixtures formed by three oils at different compositions, whereas the last consisted of a recombined live oil and supercritical CO2. Predicted viscosities are in better agreement with experimental data than those estimated from Shu’s correlation. The total mean absolute percentage error was 6.66 and 19.32 for the proposed and the Shu correlation, respectively.