The sudden occurrence of dust storms results in significant economic damage, with additional negative impacts on public health and welfare. This study investigates one of the most vulnerable areas of the world to dust storms (Ahvaz, Iran) to determine whether there are any meteorological parameters with predictive skill through which weather forecasters can confidently warn the public about the likelihood of an impending dust storm the following day. To this end, this study focuses on data including meteorological parameters, visibility and particulate matter mass concentrations for both dust event days and preceding days for the period between 2008 and 2016. Data were obtained for four monitoring stations (Naderi, Havashenasi, Edareh Kol and Behdasht) from the Iran Meteorological Administration and Khuzestan Environmental Protection Organization. Pearson correlation coefficients were used to identify influential parameters for dust storm prediction, and an artificial neural network (ANN) approach was applied to predict the maximum dust concentration. Minimum dew point temperature 1 day prior to dust occurrences showed a significant correlation ( p -value < 0.01) with the maximum 3-h mean PM 10 concentration during dusty days. A less significant relationship ( p -value = 0.045) was found when using the minimum dew point temperature from 2 days before dust occurrences. Using the minimum dew point temperature from 1 day before dust events with ANN resulted in strong forecasting skill for the maximum 3-h mean PM 10 concentration during dusty days ( R 2  = 0.71). Therefore, dew point temperature may provide predictive skill for the next day’s dust events.