Models were developed to assess the risk and predict the severity of gray leaf spot of maize, and to describe relationships between environmental variables and the rate of lesion expansion and sporulation of the causal organism, Cercospora zeae-maydis. Environmental, genotype, and site-specific data were collected from 50 locations in Iowa between 1998 and 2002 and used as input variables for model development, while disease severity at the R4/R5 growth stage of maize was used as the response variable. Pre-planting data were used to develop risk assessment models using ordinal logistic regression and classification and regression tree (CART) modeling approaches. The logistic regression models correctly classified 66 to 73% of the validation cases, while the CART model correctly classified 56 to 73% of these cases. All-subsets regression and artificial neural network (ANN) models were used to predict the severity of gray leaf spot based on early- and mid-season data. All-subsets regression was performed to select the best subsets of predictor variables based on Mallow's Cp criteria. These variables were then used as input for ANN model development. A three-layer, feed-forward, back-propagation network with three hidden nodes was used to model the data. A random sample of 60% of the cases was used to train the network, and 20% each for testing and validation. The predictive accuracy of the top four networks ranged from 70 to 75%, with mean squared errors ranging form 174.7 to 202.8. Quadratic regression was used to model the relationship between temperature and lesion expansion, and between temperature and sporulation of C. zeae-maydis at 100% RH, while loess nonparametric regression was used to model the relationship among sporulation and, temperature and relative humidity. Optimum temperatures for lesion expansion and sporulation were between 25 and 30°C. These results provide a better understanding of the effects of the environment on the development of gray leaf spot of maize. The risk assessment and prediction models may be used to develop timely and cost-effective management programs for this disease. More robust, mechanistic risk assessment models may be possible using data from controlled-environment studies on lesion expansion, sporulation, and other components of the disease cycle.