The impacts of mathematical models and associated parameters on radon (222Rn) and thoron (220Rn) exhalation rates based on in situ testing at building interior solid walls were demonstrated to improve data analysis techniques. The results showed that the heterogeneity of their activity concentrations within the measurement system was more significant for thoron than radon. The diurnal variation in indoor radon should be considered for better data quality. In conclusion, a model should be appropriately made and selected under the purposes and accuracy requirements of the exhalation test.