This paper emphasizes on mathematical and field work approach to diagnosing the environmental pollution for Indore, India. These applications are based on the time-series statistics and for three semi-industrial as well as residential areas. The generalized additive models finds as a best fit-model in terms of autocorrelation and reduction of over-dispersion. The interdisciplinary study works on the principal of pollutant source, meteorological parameters, pollutant types, emission rates and various chemical processes. Several chemical or industrial processes like iron and steel production, combustion of fossil fuels, biomass burning, thermal power plants are major polluter in most of the mega cities. © 2007 Springer Science+Business Media, LLC.