The diagnosis of a medical condition is based on the symptoms, physical examination and medical history of a patient. There have been many cases where treatment for a disease has not been carried out accurately due to lack of proper analysis and ignorance of a number of important factors. To make the essentials of good health practices accessible to everyone, a Naïve Bayes approach for Disease Diagnosis is proposed here. This service aims to help patients and doctors by guiding them to predict possible diseases using the symptoms provided by the user. Naïve Bayes classifier is used for classification of data. The system comprises of a symptom data set which is further categorized as Name, Attributes, and Record data. Based on the ranking of each symptom provided, the diagnosed result suggests the possibility weight of each disease. It also gives the detail of identified diseases including the cure, prevention and the possible treatment. An analysis of Naïve Bayes classification accuracy in the system is also studied. © Research India Publications.