Wildfires emit large quantities of air pollutants into the atmosphere. As wildfires increase in frequency, intensity, duration, and coverage area, the emissions from these fires have become a significant control issue and health hazard for residential populations, especially the vulnerable groups. A critical barrier to addressing the health impacts of air pollution caused by wildfires lies in our limited understanding of its true extent. This problem is expected to be exacerbated by additional factors such as the anticipated increase in wildfire intensity due to climate change, and the associated rise in fine particulate matter (PM_2.5) in wildfire smoke, which, according to recent toxicological studies, could be more harmful than typical ambient PM_2.5. The primary goal of our study is to develop a novel statistical framework that enables the forecasting of future emissions from active wildfires. This research aims to address the unquantified impacts of wildfire emissions and is a priority research area for many US federal agencies e.g., NIEHS, US EPA, and NOAA. The framework integrates physiochemical models of emissions and satellite observations with forecasting models based on spatial statistics and deep learning. Through the incorporation of these diverse datasets, we aim to improve the accuracy and reliability of our predictions regarding the spatio-temporal distribution of wildfire emissions. The potential human health impacts resulting from poor air quality during wildfires are also explored. By modeling the relationship between environmental exposures and disease risk, the burden of disease attributed to both short- and long-term impacts of exposure to wildfire events will be assessed.