The diffusion coefficient (D) of gases in heavy oils is an important mass transfer parameter to model and design gas injection processes for oil recovery. The pressure-decay technique (PDT) is one of the widely used experimental methods available to infer this coefficient. PDT records the declining gas phase pressure resulting from the diffusion of gas into heavy oil inside a pressure/volume/temperature (PVT) cell. Commonly, the gas phase pressure decay is modelled by use of Fick's second law along with gas-phase mass balance equations and assuming a constant diffusion coefficient. In this work, we evaluate two concentration-dependent diffusion coefficient functions, power-law and exponential. A simple history matching technique is used to estimate the apparent diffusion coefficient (D-o) and concentration dependency factor (m(a)) from pressure-decay data. Extensive application of our method to experimental pressure-decay tests shows that in addition to constant diffusion coefficients, both power-law and exponential functions are capable of predicting the experimental data. This implies that other experimental techniques are required to extract the functionality of diffusion coefficients with gas concentrations in heavy oil.