The empirical methods provide the suitable approaches for fog forecasting by considering the climate features and physical conditions that are necessary for fog formation. The purpose of this survey is to investigate and improve the Saunders empirical method to predict fog at the Mehrabad international airport station. To this aim, the theoretical foundations are implemented in an appropriate algorithm for the different atmospheric conditions. In this study, the upper air observations (soundings) at 12 UTC, for the period 2011 to 2016, are used as input data to compute fog point temperature. In addition, the METAR codes have been collected to assess the air mass effects on the model performance. The results show that the early version of Saunders method is challenged to predict fog in hazy or precipitation conditions, and the modified algorithm has played a significant role in increasing the prediction accuracy.