Today, along with the progress made in the development of remote sensing sensors, many Earth science researchers have paid special attention to this technology. Among the features of remote sensing data, we can mention things such as the collection of radiometric and geometric information, the short period of time to obtain information from the desired area, the variety of sensors in order to receive different information, etc. Among the various applications of these data, the identification and classification of dust storms in the last two decades are of considerable importance. The purpose of this research is to monitor dust in the southwestern region of Iran using Madis satellite data. Madis images from 2010-2015 were used in this research. According to the studies, it has been observed that in this field, different models have been presented in previous researches. In the first step of this research, the performance of fine dust monitoring models has been evaluated in the southwestern region of Iran. Then, the adaptive decision tree classification algorithm is presented with the investigations. Adaptability in this algorithm means that in this method the parameters are determined for the classification of each image using the features of the same image. After that, using the HYSPLIT model, the trajectory of dust movement has been predicted. In order to check the performance of classification methods, 8 MADIS images between 2010-2015 have been used. By examining the classification results of Madis images using AOD training data, it has been observed that among the previous research methods, the support vector machine classification algorithm in the southwestern region of Iran has more appropriate accuracy than other methods. Then, using the HYSPLIT model, the trajectory of dust in the southwestern region of Iran shows the foreign origin as the main source of this phenomenon in Iran.
)