ABSTRACT: In this paper, shows importance of molding a reliable, precise and accurate method which estimates disease severity of Late Bligh Disease predicting yield loss using image segmentation techniques based on Artificial Neural Networks (ANNs), forecasting and monitoring epidemics, Evaluation of crop germplasm for disease resistance, and to understand fundamental biological processes including coevolution. Late Bligh Disease assessments that are inaccurate might lead to faulty conclusions being drawn from the data, which in turn can lead to wrong actions being taken in disease management decision. In the proposed approach which is a disease-independent approach, a system modeling for Late Bligh Disease will be discussed considering the necessary element that participates in the enhanced model. Therefore, experimental results will be introduced and discussed in detail. The results show that the enhanced automatic recognition image processing model based on ANN gives fast, accurate and efficient severity estimation. More than 93.5 % average accuracy is achieved. Finally, disease severity is categories by calculating the quotient of lesion area and leaf area. The enhanced model can be potentially extended to cope with various kinds of disease and taking into account not only the infected area, but also the number of infected spots and disease degree.