An inability to complete daily tasks and changes in mood, energy, and activity level are symptoms of bipolar disorder, also known as manic-depressive illness type disorder. Serious consequences result from untreated bipolar disorder, such as elevated health care expenses, mortality, and morbidity, along with notable role impairment. It could take up to 15 years to get a diagnosis, despite the fact that there is a workable treatment option. Three common conditions—bipolar disorder, migraines, and epilepsy—have little known about their underlying pathophysiological mechanisms. It is rare for a patient to have co-morbidities between two of these conditions, but co-existence of all three is common. Here, we combine the information from the magnetic resonance imaging (MRI) scan with the random forest algorithm. These abnormalities usefulness in differentiating bipolar disorder patients from mental health issues or state-of-mind disorders defines patients based on their illness. Here, for accuracy, we employ machine learning algorithms such as CNN-MDRP (Multimodal Disease Risk Prediction) and Random Forest algorithm. We also use the attributes we gathered from the MRI to determine the risk factor and stage of the healthy patient. To obtain the output, we employ the machine learning algorithms discussed above along with a trained dataset. So, the system proposed gives adequate information about activity of patient such as Heart beat, motion, respiration and stress so that based on the real time data, the psychiatrist or pharma co-therapist can provide appropriate treatment according to the data provided. The provided system will also provide positional data so that the friends or relatives of that patient track him appropriately for an emergency and process with real time values derived from the sensors. So, the overall system provides the real time positional as well as sensory database which plays important role while treating the patient affected with the bipolar disorder.