This research aims in the design and implementation of a flexible Computational Intelligence System (CIS) for the assessment of air pollution risk, caused by PM10 particles. The area of interest includes four urban centers of Cyprus, where air pollution is a potential threat to the public health. Available data are related to hourly daily measurements for 2006, 2007 and 2008. This Soft Computing (SC) approach makes use of distinct fuzzy membership functions (FMFs) in order to estimate the extent of air pollution. The CIS has been implemented under the MATLAB platform. Some interesting results related to each city are analyzed and useful outcomes concerning the seasonality and spatiotemporal variation of the problem are presented. The effort reveals the severity of air pollution. Risk is estimated in a rather flexible manner that lends itself to the authorities in a linguistic style, enabling the proper design of prevention policies.