PurposeThis paper aims to present the most influential factors on classroom indoor PM2.5 (Particulate Matter < 2.5 µ), determining the level of PM2.5 concentration in five pre-schools located in the most densely populated district of the Tehran metropolitan area (district 6) as a case study to consider the children's exposure to air pollutants and introducing a suitable model, for the first time, to predict PM2.5 concentration changes, inside pre-schools.Design/methodology/approachIndoor and outdoor classes PM2.5 concentrations were measured using two DUSTTRAK direct-reading instruments. Additional class status information was also recorded; concurrently, urban PM2.5 concentrations and meteorological data were obtained from the fixed monitoring stations and Meteorological Organization. Then, the predicted concentrations of the indoor PM2.5, from introduced multiple linear regression model via SPSS, compared with the nearest urban air pollution monitoring stations data.FindingsThe average outdoor PM2.5 concentration (43 ± 0.32 µg m−3) was higher than the mean indoor (32 ± 0. 21 µg m−3), and both were significantly (p < 0.001) surpassing the 24-h EPA standard level. The indoor PM2.5 concentrations had the highest level in the autumn (48.7 µg m−3) and significantly correlated with the outdoor PM2.5 (r = 0.94, p < 0.001), the number of pupils, ambient temperature, wind speed, wind direction and open area of the doors and windows (p < 0.001). These parameters, as the main determinants, have led to present a 7-variable regression model, with R2 = 0.705, which can predict PM2.5 concentrations in the pre-school classes with more than 80% accuracy. It can be presumed that the penetration of outdoor PM2.5 was the main source of indoor PM2.5 concentrations.Research limitations/implicationsThis study faced several limitations, such as accessibility to classrooms, and limitations in technicians' numbers, leading to researchers monitoring indoor and outdoor PM concentrations in schools once a week. Additionally, regarding logistical limitations to using monitoring instruments in pre-schools simultaneously, correction factors by running the instruments were applied to obtain comparable measurements.Originality/valueThe author hereby declares that this submission is his own work and to the best of its knowledge it contains no materials previously published or written by another person.