Solar chimneys are often used in buildings to enhance natural ventilation; thus, accurate assessment of the ventilation effect is essential for their design. However, the prediction method available in the literature usually overpredicts the airflow rate. In the present study, a solar chimney with a variable gap-to-height ratio between 0.1 and 0.5 is used as the research object to develop an improved prediction method. Experiments and theoretical analyses were conducted to investigate the induced airflow characteristics, such as air temperature distribution, air velocity distribution, and the reverse flow phenomenon. The results show that the distribution of induced air temperature and velocity were highly nonuniform and that the optimum chimney gap-to-height ratio was estimated to be 0.4. The reverse flow phenomenon appeared at the outlet when the chimney gap was 400 mm, and the reverse flow continued to penetrate downward into the chimney as the chimney gap increased. The proposed method considers the variation in the pressure loss coefficient for various flow conditions at the chimney outlet. By comparing the predicted value and the experimental data, the proposed method was demonstrated to be valid for predicting the induced airflow rate with or without reverse flow. Moreover, the current study motivates further research on prediction method for airflow rates in solar chimneys.