Brown patch is a very serious disease on creeping bentgrass(Agrostis stolonifera). Pesticide application and culture measure are major methods now used to control the disease. In addition to the financial costs of application, certain fungicides may be harmful to humans and the environment. These researches concerns with brown patch forecasting system and mainly include: ①evaluate two weather-based empirical indices for its ability to predict the occurrence of brown patch infection episode; ②develop weather-based forecasting models for brown patch on creeping bentgrass in Beijing; ③develop a comprehensive forecasting model using weather factors and culture factors. The main results are as follows:1. With meteorological information and disease occurrence on creeping bentgrass from July 1 to August 31 in 2005 to evaluate two weather-based empirical models (the 'Fidanza' and 'Schumann' models) for its ability to predict the occurrence of brown patch (Rhizoctonia blight) infection episode. Three are two indices (E2m and E2h) in F96. For E2m, the false alarm ratio is 0.43, the probability of detection is 0.22, the critical success index is 0.19. For E2h, the false alarm ratio is 0.27, the probability of detection is 0.44, the critical success index is 0.38. For S94, the false alarm ratio is 0.20, the probability of detection is 0.67, the critical success index is 0.57. The performances of these disease indices are poor.2. A compare test of the E2m' performance with two different sources (on-site observation and regional meteorological forecasting) of the meteorological information was conducted. The result shows that correlation of meteorological information between two different sources is significant at the 0.01 level, the correlation of E2m' performance between two different sources is significant at the 0.01 level too. These demonstrate that the model performance is independent of the source of the meteorological information. Therefore, regional meteorological forecasting information can be effectively applied to develop turfgrass disease forecasting model.3. Developed a weather-based forecasting model to predict the occurrence of brown patch infection episode with meteorological and disease occurrence information in 2005. it is Y=1/[1+e~ (5.179+0 346T-0.156RH) ], in which Y is represent the probability of occurrence of brown patch infectionepisode, T is represent the minimum air temperature in 24 hours, RH is represent the mean air relative humidity in 24 hours. By evaluating its ability, the result shows that its critical success index is high and it can be used in experimental region. Minimum air temperature , mean temperature and RH all had significant correlations with disease percentage and disease index at 0.01 level, but is insufficient to4. forecast them. Disease percentage and disease index had significant correlation at 0.01 level too, and the linear regression model is Y=0.67X-1.303 (Y represent disease index, X represent disease percentage).5. Three cultural factors are chosen to test their relationships with brown patch. The results show that amount of nitrogen applied and cutting height had significant correlation with brown patch severity, while cutting frequency had little influence on it.With two culture factors combined with two meteorological factors, the disease percentage forecasting model and disease index forecasting model of brown patch on creeping bentgrass had developed.Disease percentage forecasting model is : X=-82. 195+3. 736N-3. 305H+4. 476T+0. 405RH disease index forecasting model is: Y=-9.519-1.172H+2.032N+0.72T+0.297RH.