Dusts pose serious danger to the human health and environment. A significant part of dust emissions3 comes from motor traffic. In consideration of the very harmful impact of dusts, increasingly restrictive limitations are imposed by law on the dust immissions4 in atmospheric air. Originally, the immission of particulate matter fraction PM10 (with characteristic particle dimension below 10 μm) was limited by law. Since 2010, the immission limitations have been extended to the particulate matter fraction PM2.5 (with characteristic particle dimension below 2.5 μm). The environmental risks posed by dusts are assessed on the grounds of measurements, modelling of pollutant emissions from natural sources and from sources connected with civilization, and modelling of dissemination of such pollutants. This paper presents models of the immission of specific size fractions of the particulate matter coming from automotive sources, based on functional similarity (behaviouristic models). In the PM10 immission models, a relationship is postulated according to which the PM10 particulate matter immission rises with increasing nitrogen oxides and carbon monoxide immissions. In the PM2.5 particulate matter immission models, an increasing dependence of the PM2.5 immission on the PM10 immission is postulated. Similarly, the PM1 immission is postulated to increase with rising PM2.5 and PM10 particulate matter immissions. The paper shows results of identification of the PM10 and PM2.5 immission models for two air quality monitoring stations. The model coefficients were found to be very susceptible to the conditions of pollutant emission (especially to the types of pollutant emission sources) and to the pollutant dissemination conditions.