Late blight, caused by Phytophthora infestans, is one of the most important diseases of potato in the world and Iran, especially in Golestan province. 16 models have introduced for forecasting the disease in the world, sofar. In order to developing a forecasting model, wether and disease occurrence data during the recent 10 years were used.then, 22 variables were built using daily temperature, relative humidity and percipitation data during Farvardin (March 20 to April 19). Study of these variables using logistic regression and discriminant analysis showed that four variables nTRH10, nTRHPR10, nRH10 and nTRHP20 were strongly ralated with occurrence of disease. Four forecasting models were developed for prediction of disease occurrence, based on these variables. Variable nTRH10 had the most strong relation with disease occurrence, and its mean values were 45 and 118.75 in years without and with disease, respectively. Prediction accuracy of the final model (developed based on nTRH10) was 100%.
)