The method for rapid detection of hazardous pollution of the atmosphere of cities, which is based on dynamic measures of recurrence (repeatability) of the states of the pollution concentration vector, was developed. The new scientific result is related to the use of the unconventional modification of the known measures of recurrence based on the dynamic window averaging the current recurrence of the states of atmospheric pollution concentration. One type of a window has the width that is increasing over actual time of measurements. The other type uses the window of a fixed width that is movable over the time of measurements. The modified measures take into consideration the integrated nature of explicit and hidden destabilizing factors that contribute to current pollution concentration at the point of control. In this case, it is emphasized that there is no need to take into consideration the traditional meteorological or other conditions when identifying hazardous pollution of the atmosphere. The developed method makes it possible to detect rapidly not only explicit, but also hidden dangerous pollutions of the air basin in cities and thus to improve the effectiveness and timeliness of the measures to reduce the harmful effects of pollution of the atmosphere on the population and the environment. Nitrogen dioxide was considered as a hazardous pollutant during the experimental verification of the method. It was established experimentally that the dynamics of the concentration of nitrogen dioxide in the atmosphere of a typical urban configuration has a fractal structure, which depends on the pollution control points. In this case, these structures are characterized by the existence of the elements of periodic and extreme topologies with sharp changes in dynamics. The modified measures were found to characterize the features of specific structures and to detect not only explicit, but also hidden hazards of atmosphere pollution. In this experiment, the dynamics of the modified measures varies from zero to 0.78 units. It was shown that the maximum value of the measures belongs to the interval of observation, which is determined by 12–36 counts. It was established that at the studied control points, current concentrations of nitrogen dioxide exceeded the limit concentrations by 2.75–4.5 times and admissible maximum single concentrations – by 1.3–2.1 times. It was determined that abrupt changes in the dynamics of the modified measures can serve as an indicator of not only explicit, but also hidden hazardous pollution of the atmosphere of cities