Childhood asthma exacerbation has multiple risk factors that occur concurrently in the environment - including extreme meteorological conditions, air pollution, aeroallergens, and respiratory virus infections. Few studies have predicted asthma exacerbation based on multiple time-varying environmental risk factors, together. In this study, we constructed an autoregressive integrated moving average (ARIMA) model to predict "high-risk" days for childhood asthma exacerbation in Philadelphia, PA from 2011 to 2016, during the aeroallergen season of each year, using a total of 28,540 asthma exacerbation case events identified from electronic health record (EHR) data. We selected predictors from quantile weighted sum regression (gQWS), incorporating temporal lags and season-stratification (early- vs. late-season), which were entered subsequently into multivariable ARIMA models. We found that daily nitrogen dioxide (NO