The thermal-wave resonant cavity (TWRC) technique has been used for thermal diffusivity measurements by many researchers. This study aims to reduce the uncertainty associated with TWRC signal processing (curve fitting) by means of numerical simulation and experimental verification. Simulations show that the plot of signal amplitude versus cavity length can be fitted to a simplified model reported previously when the initial fitting position is at least twice the thermal-wave diffusion length (2 μg), and that the uncertainty caused by different end positions is negligible in the range of 6-10 μg. Upon consideration of the simulation results, signal-to-noise ratio, and clearly defined amplitude curve shape, fitting ranges of about 2.2-8.0 μg and 2.2-8.7 μg were chosen for the experimental data. Thermal diffusivity values (1.438 ± 0.001) × 10-7 and (1.436 ± 0.001) × 10-7 m2 s-1, respectively, were obtained for distilled water, in excellent agreement with the accepted literature value. The ratio of standard deviation to the mean value is smaller than 0.07%, one order of magnitude lower than typical results reported in the literature. Similar simulation results were obtained for air and methanol as intra-cavity samples.