Heat capacities for single-phase mixtures of the natural gas components methane (1), propane (2) and n-heptane (3) have been determined at temperatures (197 to 421) K and pressures up to 32 MPa using two differential scanning calorimeters with a combined standard uncertainty of (2.0 to 2.6) % (k = 1). In addition, measurements were performed at pressures (0.01 to 4.40) MPa higher than saturation conditions to estimate the heat capacities of the mixtures at their bubble points. The ternary mixture data were compared with three models: the Groupe Européen de Recherches Gazières (GERG) 2008 multi-parameter equation of state (EOS), the Peng-Robinson (PR) EOS, and the Statistical Associating Fluid Theory (SAFT)-γ Mie EOS incorporating group contributions. The relative deviations of the measured heat capacities from the values calculated by the three models show similar, systematic dependences on density, with larger deviations at cryogenic temperatures. The root mean square deviations from the measurements (with ideal gas heat capacity corrected) were 8.2%, 7.0% and 5.9% for the GERG-2008, PR and SAFT-γ Mie EOS, respectively. The presence of n-heptane increased the deviation up to 20% at the lowest temperature. The addition of methane to the binary mixture [0.500 C3H8 + 0.500 C7H16] was found to always increase the heat capacity. This work shows that the SAFT-γ Mie EOS can describe the single-phase measurements at above-ambient temperatures, while none of the models provides reliable predictions for cryogenic single-phase and near-bubble-point measurements.