Satellite data have tremendous potential for mapping the global distribution of aerosols and their properties. The MODIS on NASA's Earth Observation System (EOS) Terra and Aqua provide near daily observations of aerosols over global land and ocean surfaces with moderate spatial resolution. Recent studies have demonstrated the potential of using satellite-derived columnar aerosol properties as a surrogate for fine particulate matter (PM2.5) air quality over local, regional and global scales. However, several outstanding issues remain in using satellite data because most satellite data provide column information (aerosol optical depth: AOD) whereas air pollution near the ground is the most important parameter (PM.25) affecting human health. These two measures of pollution correlate well with each other but their relationship is highly dependent on local meteorological conditions as well as vertical distribution of aerosols. The current study will explore the possibility of combining satellite data with model derived meteorological field into two different statistical model frameworks. The first approach will use simple multi-variable regression analysis where meteorological parameters along with satellite data will be used to model daily mean surface level PM2.5 mass concentration. In the second approach, an artificial neural network system will be used using same data sets to estimate PM2.5 mass concentration.