Despite advances in medical sciences, many patients cannot gain benefits from developed Methods, especially health care specialists in areas such as, health detection systems are not able to address the problem of high mortality due to lack of proper systematic detection is increasing the death rates in the patients. Brain hemorrhage is one of the most frequent forms of hemorrhages among the people all over the world. The social problem of this research addressed is the lack of system for detection and Classification of brain hemorrhages. This work explores the possibility of finding a solution in three phases. The results of Neural Networks in medical diagnosis and classification computerized algorithms for diagnosing brain hemorrhage by examining Computer Tomography images of patients. The results indicate that the system uses a classification approach and has an accuracy of 90% classification evaluated by domain experts, recommended medical referral decision.The results of this work at various phases could study potentially useful to develop further detection and classification system for medical community in terms of predicting patients who are likely to have brain hemorrhages.